Robotics brings together a wide range of different machines including Pepper partnering with soft-bank; the Boston Dynamics humanoid robot Atlas, which can do backflips in movies and television and a plethora of humanoids and Bots that leave the human mind with awe and inspiration to achieve new tech heights. Much that the technology that powers robotics continues to achieve new pinnacle; people not familiar with the developments tend to hold polarized views, ranging from unrealistically high expectations of robots with human-level intelligence, or an underestimation of the potential of new research and technologies.
Over the past years, questions have been asked about what is actually going on in deep reinforcement learning and robotics industry. How are AI-enabled robots different from traditional ones and their underlying potential to revolutionize various industries, what is the new excitement the robotics industry holds for the future.
These questions point towards the challenging world of robotics and how difficult it can go to understand the current technological progress and industry landscape, to enable tech giants and newbies alike to make predictions for the future.
The Uniqueness Behind the AI powered Robots
So what is about the robot evolution from the automation to autonomy? What started off as a quest to make routine work easy through automation has come a long way towards full robot autonomy?
AI brings a game changer approach to robotics by enabling a move away from automation to true self-directed autonomy. When the robot needs to handle several tasks, or respond to humans or changes in the environment, it essentially needs certain levels of autonomy. The path from autonomy has been an uphill but a truly worthwhile change. According to a source, the evolution of robots can be explained by burrowing case studies from the autonomous car space. For an easy explanation of the process underlined below, robots are defined as the programmable machines capable of carrying out complex actions automatically.
• Level 0 stage is also called as the No automation stage where people operate machines, there is no automation without any robotic involvement.
• Level 1 stage is the driver assistance level, where a single function or task is automated, but the robot does not necessarily use information about the environment. Traditionally, robots are deployed in automotive or manufacturing industries programmed to repeatedly perform specific tasks with a high precision and speed.
• Level 2 stands for partial automation where a machine assists with certain functions, using sensory input from the environment to automate some operational decisions. Examples include identifying and handling different objects with a robotic vision sensor. In this stage, robots lack the ability to deal with surprises, new objects or changes.
• Level 3 is the Conditional autonomy where the machine controls the entire environment monitoring, but still requires a human’s intervention and attention for unpredictable events.
• Level 4 is the high autonomy stage where the machine is fully autonomous in certain situations or defined areas.
• Level 5 is the complete autonomy level powering the machine with full automation in all situations.
The Current Stage of Automation
Today, a majority of robots deployed in factories are non-feedback controlled, or open-looped implying that their actions are independent from sensor feedback as that happens in level 1 stage as discussed above.
Few robots in the business act and take commands based on sensor feedback as that happens in Level 2. A collaborative robot, or co-bot, is designed to be more versatile empowered to work with humans; however, the trade-off is less powerful and happens at lower speeds, especially when compared to industrial robots. Though a co-bot is relatively easier to program, it is not necessarily autonomous to handle. There is often a need of human workers to handhold a co-bot every whenever there is any change in the environment or the task.
Pilot projects integrated with AI-enabled robots, have started to become a regular feature incorporating a Level 3 or 4 autonomy, like warehouse piece-picking. Traditional computer vision cannot handle a wide variety of objects like that in e-commerce because each robot needs to be programmed beforehand and each item needs to be registered. However reinforcement learning and deep learning has enabled robots to learn to handle different objects with minimum human assistance.
In the times to come, there might be some goods that robots have never encountered before which would need a support system and a demonstration from human workers bringing the level 3 of automation. In the times to come, improvements will be seen into algorithms to get closer to full autonomy as the robots collect more data and improve through trial and error in Level 4.
Taking a clue from the autonomous car industry, robotics startups are additionally taking different approaches to autonomy for their robots. Some aspects believe in a collaborative future between robots and humans, and focus on Level 3 mastery. While in a fully autonomous future, skipping Level 3 and focusing on Level 4, and eventually on Level 5 will be difficult to assess the actual level of autonomy.
The Age of AI-Enabled Robots in Industries
Taking the brighter side, robots are being used in a lot more use cases and industries than ever before. AI-enabled robots are running warehouses, in a semi-controlled environment, picking up critical pieces that are fault-tolerant tasks. On the other hand, autonomous home or surgical robots will be a reality of the future, as there are uncertainties in the operating environment, where some tasks are not recoverable. With the change in time, the human eyes will see more AI-enabled robots being used across industries and scenarios as reliability and technology precision improves.
The world has seen only about 3 million robots, most of which work on welding, assembly and handling tasks. There have been very few robot arms being used in varied industries like agriculture, industries or warehouses apart from electronics and automotive units, due to the limitation of computer vision.
Over the next 20 years, the world will witness an explosive growth and a changing industry landscape which will bought by the next-generation robots as reinforcement learning, cloud computing and deep learning unlock the robotic potential.