Robotics and Autonomy

Part
01
of five
Part
01

Aviation Industry Insights - Robotics and Autonomy

Autonomy and robotics are used widely in the aviation industry, especially for defense and civilian purposes. Autonomous systems are being studied and advanced concerning drones and even traditional warplanes; robotic dexterity allows AI systems to control an actual airplane cockpit like a human pilot. Autonomy is also important in the operations of airports around the world. These points and more are discussed below.

Autonomous systems

Manned-unmanned autonomy

  • Thanks to new funding, new manned-unmanned projects make up 8% of all new projects in the different parts of the military, including the Air Force.
  • The Army is now focusing on solving technical problems when controlling vehicles remotely. The new systems are set to help controllers make decisions quicker than fully-manned missions.

Collaborative autonomy

  • DARPA has developed a system called CODE (Collaborative Operations in Denied Environment) that leverages collaborative autonomy to navigate drones that can find, track, identify, and engage targets without much human involvement.
  • Before, UAVs would revert to the last programmed mission if operators lost communication, but with CODE, multiple teams of systems can share information and collaborate to adapt and respond to threats wherever they are.

Trusted autonomy

  • Transparency is very important in trusted autonomy as it influences trust. If a human can't understand why a machine made an action and vice versa, then the whole system falls apart.
  • When engineers are designing an autonomous system, they have to understand what will make a user trust it.
  • Researchers at an Australian University are building and testing a system that adapts to the cognitive needs of human operators by determining their state of fatigue, stress levels, attention, workload, and level of trust. This system has neuro-physical sensors that allow the machines to infer the cognitive state of the human operators and adapt.

Autonomous vehicles

  • At an airport in Sweden, autonomous vehicles were tested for the first time to do tasks at the airport such as clearing snow to keep runway edge lights clear. This prevents runways from getting closed down due to 15% or more of the lights being non-operational.

Robotic perception

  • Even though robotic perception is not yet up to human perception, robots are currently being used in airports around the world to help passengers navigate widespread terminals and to answer pressing questions.
  • One of these robots that is being used and developed right now is Spencer (Social situation-aware PErceptioN and action for CognitivE Robots). Because of perception, the robot was able to scan boarding passes and adjust its speed to that of the group while guiding a group of passengers to their departure gate and avoiding obstacles.

Robot dexterity

  • A Carnegie Mellon spin-off company, RE2 Robotics, which has experience in building bomb disposal robots, was commissioned by the United States Air Force to develop the Common Aircraft Retrofit for Novel Autonomous Control (CARNAC) system. This system is being used to replace real pilots for unmanned operations.
  • CARNAC has been successful so far as a system has been built that is capable of takeoff, normal flight, responding to emergencies, and landing.
  • The Air Force has been involved in other robotic flight projects, such as the ROBOpilot project, which has been able to make a successful flight of a civilian airplane over Utah. The ROBOpilot works like a human by physically taking the yoke and flipping switches and pressing buttons.

Grasping and manipulation

  • Researchers are currently developing micro aerial vehicles (MAVs) that use grasping and manipulation methods to perch and hold on to objects.
  • In one experiment, researchers went against the common methodology of using a fixed reference frame to identify objects and used a method called visual servoing. This is a better method as it helps with objects that are difficult to grasp such as cylinders and moving objects, even though much more research is needed where that is concerned.

AI and robotics for the aviation industry

Part
02
of five
Part
02

Energy Industry Insights - Robotics and Autonomy

Currently, autonomous and robotic systems are being used to help decommission nuclear facilities, inspect oil and gas platforms and pipelines, inspect wind turbines, and other tasks. These systems help to do very dangerous jobs without the risk of human life, while making them more efficient and less costly. These points are discussed in greater detail below.

Autonomous systems

Manned-unmanned autonomy

  • PricewaterhouseCoopers (PwC) has estimated the value of unmanned aerial systems or drones in the world's power and utility sector to be $9.46 billion.
  • The power and utilities sectors are taking a large interest in drones and unmanned aerial systems (UAS) because they cut operation and maintenance costs and boost worker safety when work will have to be done in potentially dangerous areas.
  • These AI-powered systems are being used to find cracks in faulty solar-paneling to detecting texture changes and cracks in steel components of nuclear power plants.

Collaborative autonomy

  • The technology is now present that allows energy experts to sit in collaboration rooms and remotely operate oil and gas facilities hundreds or even thousands of miles away in the Gulf of Mexico. This saves on traveling time and expenses and helps companies shift some focus to other needs such as job rotations, maintenance safety, personnel changes, and government regulations.
  • Collaborative autonomy will also help to boost cyber systems' resilience to protect against attacks on the power sector.

Trusted autonomy

  • GE is currently a leader in trusted autonomy research as they make and use trusted systems for power plants and wind farms. They are currently developing an autonomy stack that uses reactive planning to decide what to do at a given moment. This technology will help power plants and wind farms to make inspections more efficient and quicker with the use of multiple inspection vehicles.

Autonomous vehicles

  • Assuming the autonomous vehicles of the future will all be electric and that the trend of declining grid loads continues due to energy efficiency, it is predicted that utility companies will look to using autonomous vehicles to increase their demand and balance loads.

Robotic perception

  • The General Robotics, Automation, Sensing, and Perception (GRASP) Lab at the University of Pennsylvania School of Engineering and Applied Science helped a team led by the Southwest Research Institute to develop an unmanned aerial system to fly into containment vessels of the damaged units at the Fukushima Daiichi nuclear power plants in Japan. These robots will autonomously operate within the area to find damages in a highly radioactive area.
  • This is an improvement upon the earlier robots that swam through the sunken reactors. These robots only had cameras that sent pictures which a team of humans had to manually analyze.

Robot dexterity

  • Robotic arms are now being developed and used in scenarios that are unsafe for humans. Robotic arms developed by a Lawrenceville-based company, RE2 Robotics, have been used at the Fukushima site since January 2019 to clean up debris. This system is autonomous.

Grasping and manipulation

  • The robotic arms developed by Shadow Robot Company from London are partially controlled by humans, where the robot's grasping mimics that of the human operator. They are slightly haptic and send touch signals to the human operator, letting them feel what the object feels like. They are being developed to be used in the nuclear decommissioning process.
  • This follows an advanced teleoperation set-up, where the human operator can feel the amount of force that is being applied and the distance at which the gripper is opening and closing to grasp an object.

AI and robotics for the energy industry

  • Advances in AI and robotics research have helped separate humans from installing oil and gas equipment, inspecting and repairing power lines, maintaining underwater pipelines and wind turbines, and decommissioning nuclear plants.
  • Currently, an AI-enabled robot is being tested at Total's Alwyn platform, which is 273 miles northeast of Aberdeen, Scotland. This system is being used to carry out visual inspections and detect gas leaks. Even though this program is successful, experts believe it will take at least five years before the use of these robots become commonplace.
  • Even though work will be run by remote operations centers which will make some jobs redundant, it will spur new job classes such as data scientists, statisticians, and machine-learning specialists in the oil and gas industry.

Part
03
of five
Part
03

Healthcare Industry Insights - Robotics and Autonomy

The healthcare industry is currently being helped by robots, especially in the operating room, as robotic arms are being used in minimally invasive procedures and some other risky procedures. With the introduction of AI, ML, and other systems, healthcare costs are set to decrease along with accidents and missed appointments. These points and more are discussed below in greater detail.

Autonomous systems

Manned-unmanned autonomy

  • Recently, the US Army Medical Research and Materiel Command outlined multidomain battlefield operational concepts which will define ground combat operations between 2025 and 2040.
  • In this plan, manned-unmanned teaming (MUM-T) will be leveraged to provide complex acute and critical care and provide support in constrained or denied environments.
  • These systems will be able to deliver blood products and provide telehealth services when prolonged care is needed but evacuation is not possible.

Collaborative autonomy

  • Even though machine learning has helped in disease diagnosis, not much is done with interventional care. This is not good, as 400,000 deaths occur annually due to medical error in the surgery room.
  • Seventeen percent of those deaths are preventable.
  • Minimally invasive robotic surgery (MIRS) helps somewhat, but as the robots are fully controlled by man, the problems still exist. With a new type of collaborative autonomy, surgical systems based on machine learning, advanced visualization, and data analytics will help surgeons through surgical decision support, process optimization, and assessment and feedback.

Trusted autonomy

  • Modern operating rooms are moving from a Robot-assisted Surgery (RAS) model that uses teleoperating tools to an agent that is able to make its own decisions and goals by way of trusted autonomy.
  • As the trust grows, these systems will no longer have to depend on the dexterity and hand-eye coordination of the surgeon. These trusted systems will take over where the surgeon falls short, and it will do so transparently.

Autonomous vehicles

  • Autonomous vehicles will provide a lot of value as the huge problems of missed appointments and costs will be reduced. Currently, there is a $150 billion cost to the healthcare sector based on all missed opportunities.
  • Annually, 3.6 million outpatient appointments are missed due to transportation.
  • There are seven major cities around the world are participating in the Bloomberg Aspen Initiative to develop policy, technology, means, and methods to how autonomous patient delivery will be done in the future.
  • CBRE Research estimates that by 2029 patients will be accustomed to the idea of being picked-up by an autonomous car and brought to an appointment.

Robotic perception

  • Through perception augmentation, computers aid surgeons to navigate through different parts of the body to improve surgical plan execution and avoid accidental trauma to any vital organs. Imaging is currently done through the use of ultrasound arrays and Computed Tomography (CT) scanning.

Robot dexterity

  • Harvard researchers have recently created a hybrid rigid-soft robotic arm. This arm is attached to an endoscope and is able to stay flat for the time the endoscope slithers through the body, and pops up when it arrives at its destination. The arm is then able to help in high-risk surgical procedures. Because these devices can be scaled down to as small as one millimeter, they can be used in lung and brain procedures.

Grasping and manipulation

AI and robotics for the aviation industry

  • Biorobotics teams are currently developing snake-like robots that are used in minimally invasive surgery to delivery therapies and perform diagnostics.
Part
04
of five
Part
04

Manufacturing/Inspection Industry Insights - Robotics and Autonomy

Robotics has long been a part of the manufacturing industry. However, with machine-learning, new vision sensors, and other advancements, manufacturers are more willing to invest in autonomous systems. They have also been shown to be a good return on investment. These points and more are discussed in further detail below.

Autonomous systems

Manned-unmanned autonomy

Collaborative autonomy

Trusted autonomy

  • In a human-machine partnership, the machine must behave in an intuitive and explainable way from the perspective of the human. This is essential in deploying autonomous systems from the laboratory into production in manufacturing assembly environments. Experts suggest doing this through social planning, where the machine maintains and reasons with an explicit model of the goals, intentions, and beliefs of the humans with which it interacts.
  • An example of this is where a robot shares a task with a human during assembly. The robot must be able to adjust to the preferences of the human operators, such as the order in which the human wants to put together the parts.

Autonomous vehicles

  • Nine percent of manufacturers have adopted some sort of semi-autonomous or autonomous mobility within their operations and 11% are expected to do so in the next three years, according to a study conducted by PwC. They caution that these numbers must not be considered as low, as 30% will be considered as mainstream adoption.
  • Manufacturers are currently waiting for there to be a good return on investment before they jump into autonomous mobility. 86% of manufacturers say that a lower cost will make them consider investing and 60% say that cost was the top barrier to adoption.

Robotic perception

  • Robotic perception and planning are based on machine learning-based algorithms that sometimes form a trust problem as decisions can be made on deep learning systems that rely on large quantities of training data. This data can be good and bad, which makes the best practices for data collection hard to determine.

Robot dexterity

Grasping and manipulation

  • Algorithms are being developed for robotic motion, where robotic arms are taught to grasp objects in cluttered environments and manipulate malleable and deformable objects.
  • Instead of representing deformable objects as absolutes that must be put in the algorithm as multiple parameters, experts are instead looking into representing objects based on distance constraints and coming up with control and planning methods based on this. This would enable these robots to grasp many objects and know whether to crush, fold, gently handle, or carry out any other action to an object.

AI and robotics for the manufacturing industry

  • Just like there are snake-like robots that can do search and rescue missions in tight spots such as fallen buildings, experts envision that there will be robots that are used to inspect tight spots in the manufacturing of wings or boats. These robots will also be able to paint inside car doors and do inspections in nuclear power plants.

Part
05
of five
Part
05

Insights and Trends - Robotics and Autonomy

There is an increase in the use of mobile piece-picking in warehouses and logistics facilities. In addition, there is also an increase in drone deployment in agriculture and health among other sectors. Below is a detailed report on current, emerging, and/or future trends in robotics and autonomy.

Mobile Piece-Picking is on the Rise

Commercial Drone Deployments are on the Rise

  • There has been an increase in the use of drones for commercial applications. In addition, businesses and research firms are testing and collecting data to explore the possibilities of using drones in more manual processes.
  • In construction, drones are being used to monitor the progress of a building against a blueprint. They are also being used in agriculture to monitor levels of moisture. Robots and unmanned aerial vehicles (UAVs) are being used because they can easily navigate through pipelines, bridges, and utility towers.
  • Drones are also being used to survey land, offer superior endurance and intelligence, facilitate communication and management, improve security, and offer transport and inspection.
  • DroneDeploy is a drone mapping app that offers the ability to automate flight plans. DroneViewTech is a business-to-business company that offers services like drone tech, aerial imaging, surveying, photogrammetry, civil engineering, and GIS.
  • Key industries using drones include construction, mining, agriculture, surveying, and real estate.

Adoption of Internet of Things on the rise

  • Going forward, robots will use smart sensors to points of production to collect data that was previously not accessible to manufacturers. This is expected to lead to increased levels of productivity and efficiency.
  • According to a report by Accenture, Industrial Internet of Things (IIoT) will contribute $14.2 trillion to the global economy by 2030.
  • The 5G network is being touted as a game changer in the adoption of Industrial Internet of Things (IIoT). This is because it allows free flow of data among interconnected devices like drones, robots, sensors, and cloud-based analytics.
  • In addition, there has been a significant increase in endpoint devices that offer even greater autonomy and process data and reduces the need to save data on thew cloud.
  • Thanks to the growing adoption, the Internet of Things industry is growing with momentum. The market was valued at $190 billion in 2018. It is growing at CAGR of 24.7% and expected to reach $1111.3 billion by 2026.
  • Leading players in IoT include Cisco, Hewlett Packard Enterprise, Google, and Apple among others.

Growth in use of Autonomous Delivery Robots

  • Over the past few years, there has been growth in the use of robotic delivery means. Amazon pioneered the trends and many companies have since joined.
  • Due to the continued shortage of labor in the supply chain market, robots will be used in many aspects such as for getting items off shelves, moving products from one point to another, and delivering products at the doorstep across the country.
  • It is estimated that the robot delivery market will grow to $34 million by 2024. In addition, research shows that by 2025, 85% of last-mile deliveries will be completed using autonomous vehicles.
  • In another report, the autonomous delivery robots space is projected to grow by over 49.5% between 2019 and 2024.
  • One of the primary factors driving demand for autonomous delivery robots is the growing presence of e-commerce players and omnichannel retailers worldwide.

Autonomous Car Market is Growing Significantly

Research Strategy

We selected trends that are happening on a wide scale, driving significant amount of change or attracting large amounts of investment. We also analyzed those trends that are registering significant amount of market growth.
Sources
Sources

From Part 05
Quotes
  • "The number of robots in warehouses, e-commerce order-fulfillment centers, and logistics facilities climbed steadily in the past year, as companies followed Amazon’s example and deployed autonomous mobile platforms. These mobile robots helped reduce the miles traveled and pounds lifted by human workers."
Quotes
  • "The holy grail for supply chain automation is mobile picking robots, according to many in the industry. However, no single solution has emerged yet, and e-commerce and shipping companies are looking at multiple routes to increased efficiency. DHL is evaluating collaborative piece picking as a way to take advantage of human and robot capabilities for small-batch, high-variability orders."
Quotes
  • "Until very recently, mobile piece-picking robots were considered a too complex, and therefore unrealistic, solution to the varying demands of the average distribution center. Now the list of limitations is quickly shrinking, as the technology becomes available to make the concept increasingly viable. The prospect of robots working alongside human workers is now very much a reality, with the scope for collaboration able to optimize warehouse processes."
Quotes
  • "The last year has been an exciting time to be in the commercial drone industry. We’ve seen drone adoption skyrocket across sectors, game-changing technology hit the market, and new partnerships form along the way. It’s also been a significant year for DroneDeploy. Our platform recently hit an important benchmark: 30 million acres mapped. In one year’s time, our global presence has tripled."
Quotes
  • "Robots will increasingly deploy smart sensors at the edge of production to collect data previously inaccessible to manufacturers. This trend is currently underway and will lead to new levels of productivity and efficiency."
Quotes
  • "The global autonomous vehicle market size is projected to be valued at $54.23 billion in 2019, and is projected to garner $556.67 billion by 2026, registering a CAGR of 39.47% from 2019 to 2026. Autonomous vehicle also known as self-driving vehicles uses artificial intelligence (AI) software, light detection & ranging (LiDAR), and RADAR sensing technology, which is further used to monitor a 60-meter range around the car and to form an active 3D map of the current environment. The vehicle is designed to travel between destinations without a human operator."
Quotes
  • "Global Autonomous Vehicles market accounted for $27.09 billion in 2017 and is expected to reach $615.02 billion by 2026 growing at a CAGR of 41.5% during the forecast period."