User:Sameeerg/sandbox

Robotics: Hardware, Software, and Impact

We have created a repository related to robotics. We have included information about articles, common industry platforms/programs, glossary, and ways to interact with robotics daily. All our members are passionate about robotics and have dedicated significant time to various robotics teams during high school. We enjoyed all aspects of robotics such as hardware, software, electrical, and learning about how our work in the robotics teams is applicable in the actual industry. At some point in our professional careers, we all want to work in robotics as software, hardware, or automation engineers.

Our goal with this artifact was to inspire more people about robotics. Even if they don’t want to pursue careers in the field of robotics, we want to inspire them to keep learning about the technologies and work being done in this rapidly evolving field. Our repository is not only targeted towards the people getting newly involved with robotics but also to people who have had experience with robotics. While compiling this information from various internet sources, we discovered new topics, platforms, and technologies previously unfamiliar to us. The robotics field is vast, encompassing both software and hardware domains, with limitless opportunities for learning and exploration. In our page, we aimed to provide comprehensive knowledge across various subfields in robotics to give readers a holistic understanding of the subject. Additionally, we wanted to cater to individuals with diverse expertise levels, ensuring that those proficient in software could also gain insights into the mechanical and electrical aspects of robotics.

Our structure currently encompasses the Mechanics, Hardware, Software, Education/Competition, Applications, and Research/News articles about Robotics. We chose this specific structure as we wanted to lay a foundation by introducing the mechanics involved in robotics before delving into the assembly of hardware components. Subsequently, we explore the software platforms and libraries essential for making hardware functional in robotics. After getting a good overview and foundation in the technical components of robotics, we then highlight current robotics competitions, spanning from elementary school to college level, offering opportunities for enthusiasts to engage and pursue their interests. We then dive into some of the myriad applications of robotics across various industries to show the current and future impact of robotics. Finally, we conclude by presenting news articles and research findings in robotics, allowing readers to learn about the latest breakthroughs in the field.

We also provided lots of links to the actual research papers, informative websites, and news articles so that curious readers can feed their curiosity even more if they are not satisfied. We also included lots of images of robots and their various components so users can be more engaged with what they’re reading.

We also want to make sure that our website can continue to be improved in the future. After all, that is a hallmark of websites - that they can always continue to change and adapt with both new information and addition of more information, along with making the website look better in general. The first obvious way we can update the website is to add more information about the topics we presented as they come up - especially in terms of the hardware and software side. We can add more terms to the glossary so that viewers of the website are able to follow content on our website with much more ease, and so they really understand the different parts of robots and why our website is constructed the way it is.

Another part we can integrate into our website is that we can move it to a different platform than Wiki and make the site in general look better. This doesn’t even have to be a website through another host website - it could be our own website that we customize with different tabs and graphics to make it interesting and appealing to any user or viewer.

We can also add more topics than we currently have - for example, we could have more hardware components and software platforms that really give an in depth overview of robotics for the user. If we were to put this on a website of our own, we would have the chance to put a lot more images, videos, and other content to enhance the experience of the user. We do think a lot of our website correlates with its other components well, however, and think it looks like it’s in good shape content wise. But hosting it on our own website could also definitely help us get more reach in terms of users.

Overall, we think this group project definitely went very well and was successful, providing holistic and comprehensive knowledge - but there are for sure some ways we can improve it.

Mechanics

Computer-aided design (CAD)

Solidworks CAD Model

To design the physical mechanisms of their devices, roboticists need CAD (Computer Aided Design) software to model parts and assemblies before manufacturing them. Solidworks is an example of a parametric modeling CAD software. This means that modeled objects are defined by dimensions, constraints, and relationships in a structured manner rather than directly manipulating features. Solidworks is widely used across many industries for its powerful design, simulation, and manufacturing tools. Onshape is another example of CAD software. It is most notable for being fully online, meaning you can use and access Onshape through your browser. The online platform also allows for simple collaboration as multiple users can open and edit the same files at the same time. Another feature that differentiates Onshape from its fellow CAD softwares is the FeatureScript programming language. Users can use and create their own scripts to efficiently design parts and assemblies to parametrize more complex features and speed up the design process.

Manufacturing

3-D Printing

3D Printing is an additive manufacturing technique that constructs physical parts from a CAD model by joining together material. The flexibility and relative speed of 3D printing has revolutionized the prototyping process as roboticists can print out almost any shape in a variety of materials in a matter of hours. The most common type of 3D printing is FDM (Fusion Deposition Modeling). The printer melts and extrudes the print material, which is typically a plastic like ABS or PLA, layer by layer until the part is complete. Other techniques like SLA (Stereolithography), which uses a laser to cure liquid resin into the desired shape, and SLS (Selective Laser Sintering), which uses a laser to fuse powdered material, can also be used for higher-end production. 3D printing technology has rapidly accelerated in the last decade as it has become readily available for the average consumer, while industrial-grade printers have been developing at a swift rate.

Laser Cutting

Laser cutter in action

Laser Cutting is a manufacturing process that uses a high powered laser beam to cut various kinds of materials including wood, plastics, and metal. Laser cutting is often used for rapid prototyping as it is a fast and low cost manufacturing process. It can be used to make a quick, physical mockup of a design in the real world. A laser cutter moves the laser around according to computer instructions and a virtual drawing to create a path to cut the material. There are different kinds of laser cutters, and some can cut more materials than others. Diode laser cutters are the cheapest and most simple, but are limited to only cutting thin wood and paper. CO2 laser cutters are more advanced and can cut thicker wood and plastics, but cannot cut metal. Fiber laser cutters are expensive, but they can cut metal. These kinds of laser cutters are often used industrially and not for hobbyists.

Glossary

Actuator

The actuator is one of the most crucial parts of any robotic assembly. An actuator is a machine that transforms energy into motion; most actuators generate either linear or rotary motion. Rotary actuators are defined by torque, whereas linear actuators are defined by force. Actuators come in a variety of forms, but the three most used kinds are electric, pneumatic, and hydraulic. Hydraulic actuators use compressed oil to generate motion. They have a very high force generation capacity and are most frequently found in heavy machines. Actuators that are pneumatic and hydraulic share many similarities. Compressed oil is used by hydraulic actuators to generate motion. Electric actuators utilize magnets and electric current. Examples of actuator applications include industrial robots, which use actuators at each joint for precise positioning and object manipulation, humanoid robots, which rely on actuators to control their limbs, joints, and overall mobility, and robotic prosthetic limbs, which use actuators to assist or augment human movement.

End Effector

An End Effector is the final link of a robot which interfaces with the environment. Typically an end effector is attached to the end of a robot arm that are able to complete the tasks that the robot is designed to do. The job of the rest of the robot is to move the end effector to the place it needs to be and then the end effector should be able to do the rest. There are hundreds of types of end effectors ont he market as it will drastically vary depending on the intended functionality. Some common types of end effectors include grippers and power tools. However, there can also be passive end effectors. End effectors equipped with sensors allow the robot to collect high amounts of data, allowing the environment to dictate how it responds. When designing an end effector, its important for roboticists to consider all of these factors.

Kinematics and Inverse Kinematics

Kinematics is the study of how bodies or systems move without taking into account the forces that create the motion. In robotics, kinematics refers to the analytical study of the motion of a robot manipulator or mechanism. Forward kinematics is the technique of estimating the location and orientation of an end-effector (such as a robot's gripper or tool) based on the robot's joint parameters (angles or displacements). Inverse Kinematics is the mathematical method of computing the joint parameters (angles or displacements) required to position and align a robot's end-effector in the workspace. In other words, forward kinematics computes the end-effector pose from given joint parameters, while inverse kinematics determines the joint parameters needed to achieve a desired end-effector pose. Inverse kinematics is an essential operation in robotics for tasks such as moving a tool along a specified path, manipulating objects, or positioning the end effector at a desired vantage point.

Degrees of Freedom (DOF)

Translation and Rotation

Degrees of Freedom refers to the amount of directions a particular robotic component can move. Degrees of freedom are often broken up into two categories: rotational and translational. Rotational degrees of freedom refer to a mechanism's ability to rotate about a fixed axis. Rotation can occur in 3 axes: roll, pitch, and yaw. Translational degrees of freedom refer to a mechanism’s ability to move along a straight line in three dimensions: the x, y, and z axes. A robot with multiple mechanisms can have redundant degrees of freedom which allow it to be more flexible and fault tolerant. More degrees of freedom allow a robot to move in more directions and reach more places. However, having more degrees of freedom often requires more mechanical complexity meaning it has more failure points. Typically, a robot designer will only design as many degrees of freedom as needed for the robot’s desired function.

Dynamics

Dynamics is the study of kinematics, the motion of bodies, and kinetics, the forces acting on bodies. Robots consist of several complex bodies with many degrees of movement. Therefore it is critical for roboticists to be able to analyze and predict these movements to a high level of accuracy. This is exactly the use of dynamics. Engineers employ countless dynamics principles to predict the motion of their robots. In addition to the motion, they need to develop an accurate estimate for all the forces acting upon the bodies in their robots. Without a calculation of the forces and torques, these devices can be either overengineered, which is a waste of resources, or under engineered, which can lead to failure. Therefore, an accurate analysis of both the kinematics and kinetics is necessary to design effective robots. Although dynamics is a mechanically-centric field, it is necessary for almost all physical applications in robotics.

Hardware

NVIDIA Jetson

The NVIDIA Jetson is an advanced embedding system that has implications and use cases in the AI field, specifically computer vision and imaging. It is a hardware tool that allows developers to harness the power of artificial intelligence through its computing power. There are several different types of NVIDIA Jetson modules with varying purposes and capabilities. The Jetson Nano was built to limit the size of the hardware device while maintaining as much compute power as possible. It is accessible and easy to use for beginners, especially. The Jetson TX2 series is optimized for power and efficiency, instead of size. It has the power to run an entire deep neural network with high accuracy which can be used in industrial and manufacturing fields. The latest Jetson Orin is ideal for the highest performance with 275 trillion operations per second. All in all, the NVIDIA Jetson platform enables developers to run AI software and programs in a production-ready setting with ease and simplicity.

Arduino

Arduino Circuit

Arduino is an open-source platform that allows you to interact with hardware and software components easily. The Arduino system is comprised of a programmable circuit board, commonly referred to as a microcontroller, and an IDE (Integrated Development Environment) software that is used to create and upload computer code to the board. The IDE software runs on your computer and can be connected to your computer simply via USB and doesn’t require a separate piece of hardware (called a programmer) which makes it easier to use and build new projects. The Arduino platform is widely popular since it allows users to build IoT projects and iterate effectively. Examples of projects that can be built on the Arduino platform include a line-following robot (which is actually what we build in CS 3630: Intro to Perception in Robots at Georgia Tech!), an Arduino Robotic Arm, a Bluetooth-controlled rover robot and numerous other innovative projects related to robotics, mechanical engineering, electrical engineering and computer science.

Software

Platforms

Robot Operating System (ROS)

ROS is in short form for the Robot Operating System. While it is not an actual operating system similar to Unix or Linux, but it is a framework and set of tools providing operating system functionality. There are 2000+ packages with different functionalities, allowing for hardware abstraction, device drivers, communication, and visualization. There are many ways to install, but Ubuntu is used most commonly. It allows for creation of gamepad teleoperation - for example, the joy package provides generic ROS drivers with joysticks and gamepads. ROS allows for support by any language using C++ wrapping, with multi platform support that allows connections between multiple device processes. It allows software writers to write software allowing a robot to function without necessarily needing to know how specifics of hardware work, abstracting it so robots can be built faster and run more efficiently. It provides ways to connect many different nodes, or networks of processes, with a central hub that can connect to multiple devices in various devices. ROS is overall a very strong functionality that continues to build upon itself.

LabVIEW

LabVIEW Diagram

LabVIEW is a graphical programming language, or technically a development environment with language G. In comparison to many other languages where words are typed, in LabVIEW visual objects are placed around a screen to make things both intuitive, easier, and faster. However, it is complex because one still needs to make sure there are no bugs, crashes, and that it can be updated frequently if needed. It is used for automated test system code development, validating designs, and testing products during manufacturing tests and things of the sort. It is also used for control and monitoring of industrial equipment and processes along with condition monitoring. Many calculations can be completed along with measuring signals, creating user interfaces, database interfaces, communication, targeting processing hardware, and for controlling actuators and motors in robots specifically. All of this makes LabVIEW a very important tool when it comes to robots and industry - something that is continuing to grow in prominence.

Computer Vision

LiDAR

LiDAR sensor

LiDAR (Light Detection and Ranging) is a method for distance measurement.  LiDAR consists of a range-measuring sensor that continuously pulses light. When the light strikes a target, it is reflected back to the sensor, which measures the time it takes for the light to return. This information is used to calculate the distance to the object. Moreover, the range-measuring sensor is usually mounted on a rotating platform, enabling the device to acquire measurements at various angles within a 360-degree circle. The sensor obtains range measurements rapidly (up to approximately 10,000 samples per second) as it rotates, providing a two-dimensional picture of the surroundings around the entire robot. LiDAR plays a crucial role in Simultaneous Localization and Mapping (SLAM) algorithms, enabling robots to map their environment and determine their location within that map simultaneously. LiDAR sensors are widely used in autonomous cars, for accurate object recognition and classification, obstacle detection, and numerous other robotic applications.

OpenCV

OpenCV is an open source computer vision library, operated by the non-profit Open Source Vision Foundation. It allows developers to build image and computer vision solutions using their own datasets while saving money, due to the accessibility of OpenCV. The algorithms and capabilities of this library includes face recognition, object detection, extracting and producing 3D models of objects, and editing an image using machine learning. Because of its ease and performance, it is one of the world’s most popular open source machine learning library today. The OpenCV library is used through Python, a high level programming language, which is known for its simplicity and beginner friendly syntax. OpenCV makes it easy to do image or video processing on your data and extract meaningful insights for users. Many businesses and corporations use OpenCV for cases like detecting brain tumors in patients, security surveillance, or autonomous drones. Essentially, OpenCV gives the ability to do image processing and object detection through code.

TensorFlow

TensorFlow is an open source machine learning tool that allows you to solve meaningful problems with a few lines of code. It was originally developed by Google Brain for Google’s internal use in 2015. Although Python is mostly used to interact with Tensorflow, the library also supports the use of other programming languages such as C++ and JavaScript. The tool prides itself on making it as easy as possible for its users to build and deploy machine learning models. Some of the possibilities with TensorFlow include building recommendation systems like the ones in Spotify and Netflix, image classification applications, and even facial recognition systems. It offers pretrained models to use, various datasets, model evaluation tools, and capabilities to bring your ML model to the production-ready level. All in all, similar to OpenCV, TensorFlow is a broad machine learning library that enables users to solve any problem they wish using the power of machine learning.

AprilTags

AprilTag compared to other fiducial markers

AprilTags are a specific type of fiducial marker for computer vision commonly used in robotics applications. Fiducial markers are points of reference used for position estimation and localization – techniques to know where the robot is. This can be used for two or three dimensional alignment with the marker. AprilTags look like QR codes, but are more simple in order to reduce processing time. AprilTags have families and each one has tradeoffs. These tradeoffs include, but are not limited to, detection range, precision, and computational complexity. The ways that these different families vary are by varying size, shape, error correction, and bit width. AprilTags are relatively easy to use as many libraries to process them are open source. Some of the techniques to process AprilTags for localization are converting the image to black and white, lowering resolution, segmentation, and lowering the adaptive threshold (classifying light and dark), and decoding the ID.

Large Language Models

Large Language models are machine learning models that are trained on, or have knowledge of large amounts of data. The power of large language models is that it can understand meaning from given text and provide apt responses, which is extremely powerful and unheard of. It does this by understand each word, how it fits with its direct context, and how it fits with the overall text, to create a cohesive picture of the input text. The state-of-the-art models like Claude 2 can work with “hundreds of pages of technical documentation or even an entire book” (AWS 2024). Because of these flexible and powerful capabilities, there is an endless number of use cases for large language models. They can be used for knowledge base answering, which means that it can answer any question using the text and documents you provide it with, and even code generation in any programming language. Large language models are truly the future of machine learning.

Glossary

PID Controller

PID Control Diagram

A PID Controller, short for Proportional-Integral-Derivative controller, is a popular feedback controller used to correct for errors. It constantly adjusts to the desired state despite mechanical errors and extraneous forces. The P-term, or proportional term, checks what the desired position is and compares it to the current position. This is the most simple to understand and implement of all of the terms. A feedback controller without I and D terms is called a P-controller. The derivative term anticipates future trends in the position by dampening possible oscillations and overshoots by the rest of the controller. The I term, or the integral term, sums up all of the past errors and then attempts to drive the errors towards zero. This is the most complex of all of the parts of a PID-controller and is often unnecessary to implement. A PID-controller needs tuning in order to operate. This could be done computationally, but in amateur applications it is often tuned manually. Basic knowledge of each of the terms is needed in order to implement correctly.

Teleoperation

The term teleoperation refers to operating a machine, or a system, from a given distance by humans. It can also be referred to as remote control, and differs from the term autonomous because humans are directly controlling the machine - in this case a robot - instead of the robot controlling itself and moving around completing tasks without direct control by humans. Teleoperation is often used in situations too complex, dangerous, or costly for humans to perform - for example, landing rovers on Mars directly controlled by humans is an example of teleoperation by robots. Other examples include drones, which are very common, along with manufacturing machinery or with bomb disposal robots. There are two parts to teleoperation - the first is a telemanipulator that a human operator uses at a distance to send signals to the robotic device that carries out the functions, which is the other part of teleoperation. It continues to be widely used around the world and will only continue to grow.

Haptic Feedback

Haptic feedback refers to how touch and vibrations can communicate feelings or sensations to a user. This is brought about by software responding to user interaction. These physical stimuli simulate tactile experiences, with examples include controllers vibrating because of certain actions or button clicking sensations provided by smartphone screens. This allows users to receive feedback and “feel” virtual objects, with this being a big integration into the field of virtual reality or VR. Additionally, this works by using technologies such as skin indentation devices, exoskeleton devices, or vibrotactile devices to slightly compress skin and imitate different sensations. In VR technology, piezoelectric actuators and linear resonant actuators allow for creating rumble and shaking sensations felt in virtual reality technology. It allows for an immersive experience, user accessibility, and touchscreen and device navigation accuracy by guiding users towards what options are best or even “correct.” It definitely continues to be something integrated in robots and many other devices and systems.

Education/Competition

VEX Robotics

VEX Robotics Competition Robot

VEX Robotics is a robotics program for elementary through university students. It allows students to learn about STEM (Science, Technology, Engineering, and Math) through various competitions. In VEX Robotics, students learn leadership, communication, presentation, programming, and mechanical design skills. VEX 123 and VEX GO are aimed at elementary students, and students learn basic programming and building. VEX IQ is the first competitive program in which elementary and middle school students compete using snap together pieces and block coding. VEX V5 is aimed towards middle and high school students and involves more complex design challenges and coding. Students are allowed limited manufacturing capabilities. VEX U is the university competition for VEX in which small university-based teams compete using the VEX V5 system and advanced code. VEX Robotics hosts the world’s largest robotics competition in Dallas, Texas in April. Across all of its programs, VEX Robotics has over 20,000 teams competing annually.

FIRST Robotics

FIRST Robotics Competition Palmetto Regional

FIRST (For Inspiration and Recognition in Science and Technology) is an international organization founded by Dean Kamen focused on building STEM education. The FIRST community has been building STEM leaders since 1989. FIRST has competitions for all ages, from FIRST LEGO League for ages 4-16, to FIRST Tech Challenge for ages 12-18, and FIRST Robotics Competition for ages 14-18. In FLL, students build an autonomous LEGO robot to complete certain tasks in a 2.5 minute match. Additionally, in the innovation challenge, the team researches a problem relating to that year’s theme and presents an innovative solution to this problem. In FTC, students build a smaller scale robot to compete with other teams in a 12’ x 12’ playing field. Finally, in FRC, students build larger scale robots to compete in high energy matches on a 27’ x 54’The FIRST program emphasizes teamwork and collaboration, even across teams to build a community of positive innovation.

RoboJackets

RoboJackets is a competitive robotics team at Georgia Tech. Its stated goals are to “promote, educate, and advance the field of robotics through diverse, in-depth projects”. It is divided into five separate teams: BattleBots, RoboNav, RoboCup, RoboRacing, and RoboWrestling. The objective of BattleBots is to create a robot that can fight other robots in an arena. RoboNav competes in a competition to mimic a Mars-style rover. RoboCup makes many small robots to play autonomous 6-on-6 soccer. RoboRacing creates an autonomous cart to race around a track. RoboWrestling creates small, fast robots in the style of Japanese Sumo tournaments. The RoboJackets also participate in outreach to the community by volunteering at various robotics tournaments throughout Georgia, hosting FRC and FTC design reviews, and creating robotics educational content. They also appear at exhibitions to talk to the community about what they do. The RoboJackets also host Georgia’s FRC Kickoff event where teams across Georgia learn the objectives for the new First Robotics Competition season.

Application

Human-Robot Interaction

As technology advances and the horizons of robotics broaden, the seamless integration of robots into society becomes all the more important. This challenge of integration is multi-faceted. On the technical side, robots need to be able to effectively communicate with humans to understand and respond to them. This may involve intricate sensing capabilities and complex learning algorithms to interpret the complex mediums of human communication. Large strides in artificial intelligence have been greatly contributing to this cause, but there is far more to go. Safety is another important criteria. When working in close proximity with humans, especially with physically laborious tasks, human safety is of top priority, often at the expense of efficiency or performance. On the other hand, there are questions about the social implications of robots as they are becoming more readily accepted in society. There is a lot to learn about how humans respond to robots and whether that differs from standard human-to-human interaction.

Medical Robotics

Laproscopic Surgery Robot

Medical Robotics are a subset of robotic automation, specialized in the healthcare and medical field. The COVID-19 pandemic accelerated the need and development of medical robotics and this field has made large technological strides over the last few years. There are a plethora of use cases for medical robotics: in rehabilitation, surgery, and pharmacy, just to name a few. In terms of rehab and therapy, robots can physically help and guide patients on their road to a complete and healthy recovery. The Lokomat is a tool that acts as a more advanced treadmill which a patient is strapped to using belt and shoulder straps. In surgery, the Da Vinci Surgical System is a robotic surgery device that uses fluorescence imaging and multi-jointed instruments. In the pharmacy space, the BD Rowa machine allows for flexible and efficient storage of medicines. This device reduces the time and need to organize packages and optimizes inputting of large amounts of medical goods.

Robotic Automation

Robotic automation, is a methodology that uses intelligent robotic tools to automate workflows in the industry. It is specifically used to take over and perform tasks that are repetitive for a human to do over large iterations. The pros of such a process include increased efficiency, increased accuracy, lower costs, faster processes. The cons are a high initial cost to get the robotic systems up and running, employees losing their jobs, and the fact that robotics are error-prone and accountability is difficult to determine in the case of an accident. Some examples of robotic automation are industrial robots in manufacturing, transporting packages and boxes, autonomous drones in agriculture, and construction robots. For example, KUKA Robotics develops industrial robotics used for manufacturing. These robots can carry a payload of up to 1300 kg, which is equivalent to approximately 2,866 pounds, and a reach of up to 3750 mm, which is about 12 feet in distance.

Social Robotics

Quori socially interactive robot

Social robotics is a wide field that continues to grow in many applications. Essentially, a social robot is an artificial intelligence - AI - system designed to interact with both humans and other robots themselves. These sociable robots can take over different job functions, can even serve as a member of the family, and can be designed with unique personalities. These can both be controlled using teleoperation but also can be autonomous systems allowing AI to control them, interacting independently based on cues and responses from the environment. These robots can provide tutoring, a telepresence, companionship, and can also provide customer engagement capabilities. While they do employ leading technology, they are not human and lack empathy, emotion, and reasoning and are still continuing to develop. They may respond unpredictably, but do their best to do what they were designed to do and can help a lot. These robots interact very well with humans and other robots and their job is to make life better in general. This is a very interesting field that will only continue to grow.

Robotic Assistants

iRobot Roomba robot vacuum

Robotics assistants also continue to be a developing field with the aim of helping humans complete certain tasks. These robot assistants assist with daily tasks such as cooking, cleaning, and pick up dry cleaning. These robots can generate reports and do other things, assisting with many tasks. A prominent robot assistant that has been around for awhile and continues to grow is the Roomba, a robot that goes around the house and cleans. It can detect objects and know where to turn - robots that can climb walls or go down stairs are currently robot assistants that can be bought. The Roomba has sensors, a bump, wheels, and brushes that work in tandem to fully clean many different types of rooms by itself, even creating a map of the house it can use to most efficiently clean the house as the human desires. The Roomba uses both infrared and photocell sensors to detect objects and clean a room effectively, getting rid of dust and other nonsense. The field of robot assistants also continues to grow and become something to make human life better in general - it is definitely a huge part of robotics.

Articles/Research

Nano-robotics

Kinesin Walking Nanorobot

Nanorobotics is a field that focuses on creating robotics at a nanometer scale. Essentially, the scale and dimensions of nanorobotics is extremely tiny juxtaposed to the traditional robots that we use. Moving on, nanorobotics is an interdisciplinary field which combines the spaces of robotics and material science. Nanorobotics is generally understood to involve the use of nanorobots, which are robots that utilize nanoscale sensors and actuators; they can be as small as 50 nm wide. For comparison, a human hair is about 80,000-100,000 nm wide. Nanorobots are used in any scenario where typical robotics are too large to operate in. For instance, nanorobots can enter the human body and deliver drugs, pills, and other molecules. These bots can interact with our DNA and various proteins in our body. For people that have certain life-threatening medical conditions, these nanobots can provide the specific molecules to their body that they need to survive with extreme precision and accuracy.

Computer Vision in Healthcare

Computer vision is a field where computers can “see” and interpret the world around them using images and videos. Through algorithms, a computer can come up with a conclusion given an input. Various processing techniques such as noise reduction, image enhancement, edge detection, and color conversions. Computer vision has various applications including object recognition, facial recognition, medical imaging analysis, quality control in manufacturing, and autonomous vehicles. In the medical field, computer vision can be used in conjunction with various imaging techniques such as x-rays, MRIs, CT scans, and PET scans. Computers can then use these images to do things like detect cancerous cells, analyze broken bones, assist in radiation therapy, and brain imaging. These techniques can recognize problems early, but are prone to error. However, in conjunction with a medical professional to verify, computer vision in medicine can lead to earlier and more efficient detection of medical conditions.

Powered Prostheses

Prosthetic arm for self-feeding

Powered prosthetic devices are an important and budding application of robotics. In the article, Gehlar et al. evaluate the current state of lower-limb powered prostheses. Unlike passive prostheses which are only able to provide some energy absorption, powered prostheses contribute net positive work. This works to help alleviate asymmetries in the users’ walking gait patterns. Additionally, powered prostheses have been shown to increase walking speed. However, there are several downsides that prevent them from being the widespread solution for amputees. First off, the added mechanical complexity greatly increases the weight. Additionally, it is far less reliable as the added functionalities make it prone to failure as well. Finally, one of the main challenges that researchers are attempting to tackle is the issue of tuning. As every patient is different, the control parameters to tune a powered prosthetic is incredibly time consuming and difficult. Researchers are working to mitigate the current limitations of powered prostheses and emphasize its benefits.

Self-Assembling Robots

Self-assembling robotics is a unique branch of robotics that has not received much public attention, but holds spectacular promise in its application. Swarm robotics is based on the principle of building modular robots that can self construct, self repair, and self reconfigure with the rest of the robot swarm. Researchers at the University of Stuttgart, Germany constructed swarms of three different types of mobile microbots to develop the processes of self-configurability and self-repair. Their theoretical model was based on a cognitive science and biological evolution model. One of the biggest challenges in developing a self-sustaining robot organism is for it to survive without human interaction. While normal robots rely on instruction or interaction with humans, these self-assembling robots were designed to act as their own organisms. Additionally, the researchers attempted to tackle the challenge of passing on functionality and building intelligence from generation to generation. This approach of building self-sustaining and evolving robots can bring about huge unexpected growth as the functionality of the robots will improve over time.

Soft Robotics

Soft Robot

Soft robotics is an emerging field of robotics which focuses on the design and construction of robots using flexible materials. Soft robotics represents a convergence of many disciplines including material science, engineering, biology, and computer science. While typical robots are made using rigid materials like metals and plastics, soft robots use materials like rubber, silicone, and fabrics. Soft robots are closely linked to living organisms as many are biologically inspired by animals like octopi and snakes. Soft robotics can refer to a specific mechanism in the robot or the entirety of one. However, there are many challenges as typical robotics components for actuation and sensing are often not flexible enough to be integrated into soft robots. Technologies that have advanced soft robotics are shape memory alloys and piezoelectric crystals for actuation components. However, soft robots offer advantages including the ability to navigate tight environments and manipulate delicate objects. This has many applications in healthcare, manufacturing, and exploration.

Boston Dynamics

Boston Dynamics "Spot" Robot Dog

Boston Dynamics is a pioneering robotics company focused on creating practical robots to tackle tough automation challenges that have never been solved before. In an interview with Yahoo Finance, Dr. Robert Playtar, CEO of Boston Dynamics, explains the current use cases revolve around industrial applications. Leveraging advanced software, cameras, and sophisticated electrical and hardware engineering, their robot “Stretch” can lift boxes and unload shipping containers autonomously. Another robot “Spot” which is four-legged can navigate warehouses and ensure safe factory operations by taking thermal measurements and reading gauges. However, to make these robots scalable and more widely available, the cost of research and development and materials need to be lowered and there needs to be a bigger market. The cost of launching one of these robots can be as high as $100 million, a figure that is expected to decrease as more companies, such as Tesla led by Elon Musk, enter the robotics automation sector. Dr. Robert believes that within this Artificial Intelligence revolution, there will be a demand for the robotic industry. He even forecasts that we might be able to see robots in our homes for daily tasks as soon as 10-20 years from now.

Self-driving cars

An example of a self-driving capable car

A self-driving car (also called autonomous or driverless car) is a vehicle that uses a combination of sensors, cameras, radar, and artificial intelligence (AI) to travel between destinations without a human operator. To qualify as fully autonomous, a self-driving car must navigate to a predetermined destination on regular roads without human intervention. The core technologies used in self-driving cars are also common in robotics. Sensors like LiDAR, radar, and other cameras are used to perceive the environment around the car. AI software process sensor data and deep learning models on driver behaviors to make logical driver decisions. Various integrated hardware and software systems are used for the control and navigation of a self-driving car. Major companies such as Audi, GM, Waymo (Google), Tesla, and Uber are all developing their own self-driving technologies and often employ engineers with prior robotics experience to lead their development. For example, Google's Waymo and Lyft collaborated to launch Waymo One, a fully autonomous commercial ride-sharing service. In addition to getting to their destination, riders can hail a self-driving car and provide Waymo feedback. If the ADS needs to be overridden, a safety driver is still included in the cars.

Humanoid robots

Ameca Generation 1 humanoid robot

Humanoid robots are a unique new invention in the realm of robotics where robots are made to look and do tasks like humans. They are being used in many ways - for hospitality, education, and healthcare. These robots can work with customers, create programming content, and measure vital signs in patients along with communicating information. Examples of humanoid robots include Atlas, part of Boston Dynamics as described above that can leap and backflip. Oceanone from Stanford Robotics Lab is a diving humanoid robot that can explore shipwrecks. T-HR3 made by car company Toyota is envisioned to help around the house and assist in childcare, while currently assisting with mobility services. Walker of Ubtech Robotics can climb stairs and balance on one leg, along with performing tasks around the house such as using a vacuum. Apollo of Apptronik can carry up to 55 pounds, greatly helping in the industries of retail and construction. The innovation of humanoid robots continues to be a hot and fast-growing field.

Natural Language Processing

Natural Language Processing (NLP) is a subfield of the larger study of Artificial Intelligence (AI). AI is essentially the use of machines to mimic the tasks and actions that humans carry out. NLP, on the other hand, studies how machines can gain deeper knowledge and understanding of language and text. It allows computers to look at text, understand what it is saying and what it means, and then do a particular action based on the text. Some common use cases of NLP include sentiment analysis and summarization. Sentiment analysis is the classification of text by its mood, tone, or sentiment. This can be used to categorize text, dialogue, or feedback as positive, negative, or neutral, for instance. Similarly, through NLP techniques, computers can summarize a large chunk of text much more efficiently than if a human had to read it. This can be used to summarize books, articles, files, and much more which improves the efficiency for the human user on the other end. NLP is an extremely powerful tool that is growing rapidly today.

Industrial Robots

Factory Automation Robotics Palettizing Bread

Industrial robots have been on the rise for many industrial and manufacturing purposes. These robots can all look very different, but many of these multipurpose machines have at least one arm that operates on three or more axes. There are currently over 3.4 million industrial robots, and that number is expected to continue to rise. There are many advantages to industrial robots - they are productive, precise, fast, and mostly safe. Disadvantages include initial cost, complex programming, and job loss. An example of these industrial robots is cartesian robots, complete with sliding joints to move objects. Articulated robots have rotary joints to move around similar to a human arm with a lot of flexibility. Delta robots have individually controlled arms connecting to a triangular base, and have lots of speed and agility. Cylindrical robots have a single, extendable rotating arm that can carry heavy loads and can help with welding and packaging materials. Overall, industrial robots are a huge innovation that will continue to rise.

International Conference on Robotics and Automation (ICRA)

The International Conference on Robotics and Automation (ICRA) is the premier annual conference organized by the IEEE Robotics and Automation Society, where leading researchers, academics, students, and industry professionals convene to discuss cutting-edge theoretical and applied research in robotics and automation. ICRA offers technical paper presentations, keynotes, tutorials, workshops, exhibitions, live robot demonstrations, and beneficial networking opportunities. It covers a wide range of topics, including robot design, control, perception, planning, manipulation, human-robot interaction, and various robotics applications. The 2023 edition, held in London, showcased innovations in manipulation, human-robot interaction, mobile robotics, and the intersection of robotics with Artificial Intelligence/Machine Learning.Key components include the presentation of real-world robotic systems and applications that are revolutionizing industries such as manufacturing, healthcare, and transportation. The conference also fosters partnerships and shares research results through published proceedings. ICRA, a distinguished conference with an 'A' rating, serves as an influential platform for advancing the state-of-the-art in this rapidly evolving interdisciplinary field of robotics.

Robotics Research

At Georgia Tech, there are a multitude of ways to get involved with robotics research through the Institute for Robotics and Intelligent Machines (IRIM). There are over a dozen labs that IRIM looks over. Major labs include the Mobile Robot Lab, Humanoid Robotics Lab, Healthcare Robotics Lab, Unmanned Aerial Vehicle Lab, and others focused on specialized areas. Robotics research at Georgia Tech covers a wide range of topics like robot design, control systems, perception, planning, manipulation, human-robot interaction, and various applications across different domains. Key focus areas include mobile robotics, humanoid robots, medical/assistive robotics, aerial robotics, manufacturing automation, agricultural robotics, and human-centered robotics. Cross-cutting themes involve artificial intelligence, machine learning, computer vision, mechatronics, and multi-robot systems. Georgia Tech is one of the global leaders in robotics research, with extensive labs, interdisciplinary academic programs, and strong industry partnerships that enable cutting-edge work across diverse areas of robotics from fundamental theory to applied systems and solutions.