Autonomous System

Systems across infrastructures have been maturing from automated to autonomous systems. The major difference being that automated systems follow preprogrammed or predefined sequences. Whereas autonomous systems, which can still operate similar to automated systems, are capable of responding to their environment and learning through the use of artificial intelligence and machine learning. This means, over time, the need for human input and interaction are reduced. Some of the more popular autonomous systems include those used in aerospace, automotive, commercial or industrial applications, transportation, and robotics.

An autonomous vehicle is any car or other passenger vehicle capable of sensing its environment and operating without human involvement or the need for a human passenger to monitor or intervene in the operation of the vehicle at any time. Furthermore, a fully autonomous vehicle would not necessarily require a passenger to be present in the vehicle at all. Autonomous vehicles are intended to drive anywhere a traditional vehicle drives and perform the same operations of an experienced human driver.

Autonomous vehicles work through sensors, actuators, complex algorithms, machine learning systems, and powerful computing systems to execute software. These sensors and the use of video cameras monitor the position of nearby vehicles, detect traffic lights, read road signs, track other vehicles, and watch for pedestrians.

The Society of Automotive Engineer (SAE) defines six levels of driving automation, ranging from Level 0 to Level 5.

Often self-driving is the phrase used to describe autonomous vehicles, whereas calling a vehicle autonomous suggests the vehicle would be self aware and capable of making its own choices. For example, if a passenger asks a vehicle to take them to work, the vehicle could instead drive the passenger to the beach. This is also why autonomous vehicles are often called automated, as self-driving more properly describes Level 3 automation, and at best Level 4, but does not describe up to Level 5 that are fully autonomous.

Some challenges faced by autonomous vehicles include the expense of systems such as lidar and the possible interference multiple lidar signals could cause for other systems. Weather conditions can obfuscate road signs, lane dividers, and change the way cameras and sensors track a car's movement across the road, while the detection of water, oil, ice, or debris in all and especially challenging weather conditions could present unique challenges to an autonomous vehicle. Furthermore, traffic conditions and laws could be challenging and could change the place autonomous vehicles occupy on the road as they share it with legacy vehicles.

State and federal regulations, especially in places where they contradict each other, also offer challenges for autonomous vehicles, as different states can have different road laws. There are a growing number of laws, especially on the state level, around autonomous vehicles; these include taxes specific to autonomous vehicles, higher emissions standards for autonomous vehicles, and the installation of panic buttons. These could impede the overall development and adoption of autonomous vehicles. Similarly, insurance questions, such as the liability in the case of an accident, provide challenge. For example, if an autonomous vehicle is considered at fault for an accident, the question becomes who is considered to be at fault—the manufacturer or the passenger.

Further, there are concerns over the ability for autonomous vehicles to accurately read and predict the behavior of human drivers. This could be especially difficult as human drivers often rely on subtle cues and non-verbal communication, such as body language and eye contact, to understand and predict the behavior of vehicles and pedestrians in a given environment. Autonomous vehicles could struggle to replicate the same kind of understanding.

Similar to autonomous vehicles, autonomous trucks offer a chance to develop a semi-truck capable of conducting all tasks a human operated semi-truck could perform. This could increase the speed of overland transport, eliminate the need for rest and comfort breaks required by human operators, and offer semi trucks that are capable of driving through all hours. Even an automated vehicle requiring human supervision provides the operator with a chance to rest and manage other tasks while driving and reduce travel times.

In the case of truly autonomous semi trucks, the vehicles and their loads could be monitored in a cloud-based operations oversight center. This would further reduce the number of humans needed to monitor the vehicles and could offer better visibility into supply chains and related delays or challenges.

Autonomous trucks are considered an easier form of autonomous vehicle to develop; similar to public transport vehicles, many of the trucks operate over planned and pre-mapped routes. This would reduce the need for a vehicle to calculate a route and navigate the route compared to an autonomous vehicle where a driver may request various destinations and detours along a driving route. The challenges of bad weather, poor traffic conditions, and detours based on construction or other road closures would still require a high level of autonomy.

In 2016, the United States Navy and the Defense Advanced Research Projects Agency (DARPA) launched the Sea Hunter, a 132-foot-long trimaran with a permanent crew of zero. This was designed by DARPA as an anti-submarine warfare continuous trail vehicle, which was intended to be an unmanned vehicle capable of traveling the oceans for months at a time with no onboard crew, searching for enemy submarines and reporting their location and findings to human operators.

These types of autonomous systems are being developed and largely deployed in the defense sector to explore future combat capabilities and the ability to develop and deploy autonomous and semi-autonomous robotic weapons in combination with crewed systems. A lot of the concepts in the automation of defense include weapons systems capable of surveying its surroundings either in a reconnaissance role to identify potential threats and targets. Further, these systems could be further automated in order to be given independence to choose to attack those targets based on a threat profile or mission objective.

Such systems would require some core elements, including a mobile combat platform, such as a drone aircraft, an autonomous ship, or a ground vehicle. Any such platform would require various types of sensors in order to scrutinize the platform's surroundings. These would be controlled by a processing system to classify and identify objects discovered by the sensors and algorithms to direct the platform to initiate attacks when an allowable target is detected.

Unmanned aerial vehicles (UAVs), also known as drones, may be the most popular or well-known version of autonomous vehicles. Used often by the military and in some civilian and research cases, UAVs can be guided autonomously or by remote control. They carry sensors, target designators, offensive ordnance, or electronic transmitters designed to interfere with or destroy enemy targets. UAVs are unencumbered by crew, life-support systems, and the design-safety requirements of manned aircraft, offering greater efficiency, substantially greater range, and greater endurance than equivalent manned systems.

Small UAVs, similar in size and appearance in some cases to civilian drones, are used by ground combat units also to extend their range of vision and offer reconnaissance capabilities.

UAVs are descended from target drones and remotely piloted vehicles employed by military forces since they debuted in the early 1980s. This came with the Israeli Defense Forces fitting small drones, which resembled model airplanes, with trainable television and infrared cameras and target designators for laser-guided munitions, all of which were downlinked to a control station.

One of the most well-known military drone or UAV is the MQ-1 Predator, which was first flown in 1994 and officially entered service in 1995. The Predator drone is powered by a piston engine, which drives a pusher propeller, capable of flying 80 miles per hour and with an endurance of 80 hours. It carries visible and infrared television, synthetic aperture radar, and passive electronic sensors and is capable of carrying anti-tank missiles.

A derivative of the Predator drone, the MQ-9 Reaper, offers improved performance and a larger ordnance load. Similarly, the Northrop Grumman RQ-4 Global Hawk is a drone platform similar to the Predator and Reaper, except it is jet-powered, capable of flying at a cruise speed of 400 miles per hour, has a 36-hour endurance, and carries a variety of photographic, radar, and electronic sensors.

Autonomous underwater vehicles (AUVs) are unmanned underwater robotic vehicles similar to the rovers NASA uses on Mars. As suggested by the name, AUVs operate independently of human input, and unlike remotely operated vehicles, AUVs have no physical connection to the operating vessel. Rather, the vehicles are programmed to know where, when, and what they are intended to do.

AUVs can carry a variety of equipment for sampling and surveying, such as cameras, sonar, and depth sensors. Unlike ROVs that transmit video via tethers almost instantaneously, AUV stores all data, including images and other sensor data, on onboard computers until it can be retrieved after the AUV is recovered at the end of a dive. AUVs can range in size, and dependent on the purpose of the platform, often can either glide on the sea surface or dive underwater and are capable of exploring at depths beyond human capability. Often these systems can stop, hover, and move like helicopters through the air, except through the ocean.

AUVs have to carry power onboard to enable propellers or thrusters to move through the water and to operate any sensors included on the platform. Most of these onboard power systems use specialized batteries, although some use fuel cells or rechargeable solar power, while others minimize energy demands by allowing gravity and buoyancy to propel them.

These vehicles can be options for ocean-based research; they can reach shallower water than boats can and deeper than human divers or tethered vehicles. Also, once underwater, AUVs are safe from bad weather and can operate for extended periods of time. Often, these are also platforms that can be scalable, modular, or both, in order to offer scientists a chance to choose specific sensors or modules depending on the research objectives. AUVs are also less expensive than other research vessels, but offer similar capabilities.

Autonomous robotics, sometimes known as autorobots or autobots, are any robotics that are capable of performing behaviors or tasks with a high degree of autonomy. These robotics are usually considered a subfield of artificial intelligence, robotics, and information engineering. The level of autonomy in robotics can give a workforce the ability to delegate dull, dangerous, or dirty tasks to the robot and offer the workforce a chance to engage in more interesting and valuable parts of the job for which they are qualified. Most robotics used in a majority of industrial applications are not autonomous, but rather programmed to perform repetitive movements, and are unable to react.

One of the most common autonomous robots is the Roomba. The little robotic vacuum is capable of making decisions and taking actions on what is in the environment, and will do its job without supervision or help. And while a consumer product, the Roomba offers many functions and capabilities that can be, and have been, translated to an industrial application.

One application for autonomous robotics is in warehouses. Many are already integrating autonomous robotics for material transport, where the robots can transport orders across a warehouse or through a shipping facility countless times. This can keep human workers free to perform more cognitive dependent tasks and leave the robots to perform redundant tasks. These redundant tasks can include order fulfillment, returns handling, raw material transport and sorting, parcel sortation, and inventory management. These robots can also be used for palletizing, where the robot can load, transport, and unload an entire pallet for a warehouse.

Similar use cases include the use of similar robotics for simple loading and unloading for transportation. This can include the use of robotics even for secure transport, in which case an autonomous robot could be fixed with a lock box and cabinet to safely transport high value materials and ensure that the proper chain of custody is followed. The use of robotics can also offer instant, accurate, and easily accessible documentation of the process.

One of the other industries where autonomous robotics already have a strong presence is in manufacturing, final assembly, and warehousing. And with the inclusion of autonomous robotics along a supply chain, these areas will see continued growth. With the maturation of robotics platforms, such as improvement in haptic sensors, these systems will increase in capabilities, allowing them to grasp objects ranging from eggs to multi-surface metal assembly parts, without the need to change programming or components.

Autonomous robots are also being used in healthcare, where they are primarily used to streamline the transport of supplies and medicine through healthcare facilities. They can also be used in infectious disease units where autonomous robotics can prevent nurses from having to come into frequent contact with potential contaminants, while continuing to ensure patients receive proper treatment. Robotics can also be used in healthcare scenarios for medical sanitation, such as carrying virus-killing UV lights or decontamination sprays to clean a room or space without exposing people to potential harm.

Similar to healthcare, autonomous robotics have been and can be used in biotechnology applications, especially for labor-intensive tasks such as sampling and maintenance of cell culture processes, constantly monitoring process inputs, and safely managing waste removal from a production line.

In swarm robotics, multiple robots collectively solve problems by forming advantageous structures and behaviors similar to the ones observed in natural systems, such as swarms of bees, birds, or fish. In order to make swarm robotics work, the robots work on specific algorithms, which are intended to imitate basic and real-world swarm behaviors. These swarms are not autonomous in individual behaviors, but rather practice a type of autonomy as a whole. This is done by adapting to changes in the environment by following specific behaviors, which can include pursuing a specific goal, aggregating or dispersing in the environment, communicating, and memorizing. Each robot has processing, communication, and sensing capabilities locally onboard and is able to interact with other robots and react to the environment autonomously.

In most swarm algorithms, individuals perform according to local rules and overall behaviors emerge organically from the interplay of the individuals of the swarm. Translated to the swarm robotics' domain, individual robots exhibit a behavior based on a local rule set that can range from a simple reactive mapping between sensor inputs and actuator outputs to elaborate local algorithms.

The industrial internet of things (IIoT) is used for smart machines and real-time analytics to make better use of data developed by industrial machines. The principal driver of IIoT is smart machines, or autonomous machines, because they can capture and analyze data in real time, and these machines can communicate findings in a manner that is simple and fast for better or more accurate decisions. IIoT is used across industries, including manufacturing, logistics, oil and gas, transportation, mining, aviation, and energy. And IIoT can be used to manage and monitor automated and autonomous machines, and coordinate autonomous machines for better coordination.

Similar to, and capable of enabling, autonomous systems, autonomous networks offer networks with minimal to no human intervention and would be able to configure, monitor, and maintain themselves independently. This could mean that the technologies could be self-provisioning, self-diagnosing, and self-healing, with advances in artificial intelligence and cloud technologies offering networks greater capabilities.

Autonomous networking is also expected to offer capabilities including connecting doctors with patients arriving on a life flight before they arrive, finding missing persons with IP cameras on a subway, selling consumers what they want when it is not in store, powering next-level robotics systems, using facial recognition for locked facilities, and enabling faster transportation of critical systems for both terrestrial applications and space-based applications.

Autonomous drive manufacturing includes the manufacturing of autonomous vehicles and their respective components to enable autonomous vehicles or other autonomous systems. This includes autonomous underwater vehicles, autonomous trucks, unmanned aerial vehicles, and drones.