Autonomous Things: Why it matters, Examples & Pitfalls

Though self-driving vehicles take the front seat when we talk about autonomous things, autonomous robots and drones can also make a difference for businesses. These technologies can lead to partial or full automation of tasks involving humans today.

Transportation, retail, security and military are some of the industries with early examples of autonomous things. Autonomy will eventually revolutionize every industry. Why is it important now? Gartner expects autonomous things to be the most important tech trend in the next years. We agree because

Technology that affects the physical world >> technology that affects the digital world. Cost of sensors(e.g., cameras) are falling The driver of autonomy, AI hardware and software is becoming more capable. Software that touches the real world rather than the digital world tends to have a disproportionate impact. Let’s consider mobile software: If we exclude social networks, the most valuable companies that rely mostly on mobile software are Uber and WeChat and they both enable activities in the real world: payments and transportation.

Improved technology for autonomy and reduced sensors costs lead to smarter autonomous things that can automate a wider variety of tasks. This results in more data and experience in autonomous things which creates a virtuous feedback cycle for them. Therefore, we expect businesses to get increasing benefits from autonomous things over time.

What are the important topics in Autonomous Things? There are four common types of Autonomous Things:

• Autonomous drones such as commercial UAVs have possible applications such as delivery vehicles (Amazon’s delivery drones) and data collectors from different environments.
• Robotics is a key driver in autonomous things, it includes smart robots, autonomous mobile robots, delivery robots, personal robots and robotic appliances. Delivery robots are already present in manufacturing, assembly, and warehousing. Robotics growth is expected to increase in supply chain networks. It helps reduce the relatively dangerous human tasks in warehouses.
• Though self-driving cars, drones and robotics are the main focus in autonomous things, with the advancements in IoT, we may start to encounter more autonomous things at the edge. Therefore AuT may shift into IoAT in the near future.
• Autonomous vehicles are the most important AuT technology in terms of economic potential:

Autonomous Vehicles

Autonomous vehicles are getting the most attention, even though there is still time before we see them on most roads due to technical and regulatory challenges. Currently, 25 countries are working on designs for autonomous vehicles and analysts estimates 8 million autonomous vehicles will be shipped in 2025. Technical components that are required for autonomous vehicles to be successful are:

• AI-based computing platform: AI is a critical technology when it comes to real-time decision making. However, AI needs a computing platform in order to make the analysis at the edge so that decisions it makes become faster.
• Deployment of 5G Vehicle-to-everything (V2X) will make it easier to develop self-driving since important information will be transmitted to the vehicle as structured data. V2X enables cars to communicate with any entity including roadside infrastructure, pedestrians and cyclists, and cloud-based services. The connection between vehicles and other entities is provided by either WLAN-based or cellular-based signals.
• Computer vision/ Sensor fusion: Computer vision uses cameras and sensors to identify objects around the vehicle. Signal fusion gathers all the information from various sensors and helps the vehicle understand the environment. Though these applications exist, they are still not robust enough. Even best-in-class applications such as Tesla’s fully self-driving vehicles have significant issues dealing with signal lights.
• High-definition maps: These maps should include the whole 3D road geometry, boundaries and allowed connections, along with semantic information about the environment, such as the position of traffic lights and road signs.

What are the current use cases? Transportation

• Robo-taxis: Tesla expects to launch robotaxis for an autonomous ride-sharing network, which may change the ownership of cars. Elon Musk said about robotaxis in 2019 Tesla Autonomy Day, “I feel very confident predicting that there will be autonomous robotaxis from Tesla next year — not in all jurisdictions because we won’t have regulatory approval everywhere.” Even though we know that Elon Musk tends to overpromise delivery time from the previous Tesla launches (Model X was launched 2 years late than expected), Tesla tends to eventually deliver. Elon Musk claims that Tesla will have more than one million robotaxis on the road next year.
• Vehicle Platooning: Peloton is using swarm intelligence, the collective intelligent behavior of decentralized systems, to build collaboration among vehicles. Platooning involves a group of self-driving trucks that move at high speed safely and efficiently because each truck communicates with other trucks and adapts its speed and braking accordingly.
• Autonomous shuttles: These vehicles navigate autonomously at sub-50km/h speeds along predetermined, learned paths. Compared to other applications of self-driving cars, autonomous shuttles are easier to regulate thanks to their restricted operations. For instance, EasyMile, an autonomous vehicle company, has deployed their autonomous shuttle EX10 in 300+ locations in more than 30 countries as of 2021. Another example of autonomous shuttles is from Mayo Clinic, a healthcare provider, they use autonomous shuttles to transport medical supplies and COVID-19 tests.