Robomorphic computing: Human tasks are now easier for robots
Robots are now going to be a part of human life. Soon, robots will handle more tasks than ever before. However, robots still lag when it comes to interaction or motion planning. Robomorphic computing could solve this problem by calculating suitable hardware for the task.
Furthermore, the use of such type of computing will enhance motion control and reaction time in robots. Currently, robots still lag while interacting with humans. They might be stuck somewhere as giving a reaction to any physical stimuli needs a lot of processing power.
In addition, a customized chip by using Robomorphic computing enables faster processing time. As a result, you will get a faster reaction time. Thus, it will essentially increase the efficiency of robots. So, they can work alongside humans in critical environments.
MIT computer scientists have come up with a new type of computing called Robomorphic computing. This computing technique can enhance the speed of robots. Traditional robots can move quickly and are powerful enough to do the necessary tasks. However, most of the time, robots become very slow, especially while interacting with humans.
In the real world, things work very quickly, and no one has time to wait for the robot to process inputs and take action. So, the robot gets stuck and doesn’t take any action until all the processing is done. So, it’s a problem of coordination between the mind and body of the robot.
Furthermore, to solve this problem, Robomorphic computing could do wonders. First, Robomorhic computing uses the robot’s physical layout and robot’s tasks or applications. Then, it generates a chip that reduces the robot’s response time for the intended task.
How does Robomorphic Computing work?:
The robot is a very complex machine. It works by coordinating software and hardware to do tasks. A robotic operation involves three main steps. Step one is data gathering from various sources of data like sensors, cameras, etc. The second step involves mapping and positioning. In this step, robots create a map of the surrounding world based on inputs. Then, they position themselves on that map. Finally, the third step is the next course of action, considering all the scenarios above, including the intentional tasks.
Furthermore, all of the above needs a lot of computing power. In the real world, the dynamic environment changes very quickly. Additionally, they have to work safely alongside humans. So, they have to think and take action very quickly. Robomorphic computing could solve this problem.
In addition, a lot of work is going on to improve the robot’s software side. However, the modern algorithms on the current CPUs are still not enough to solve this problem. Hence, scientists think we need a new set of hardware. That means going beyond standard CPU chips with hardware accelerators like GPU.
GPU is a common hardware accelerator. GPU chip has parallel processing ability. Hence, you can simultaneously process thousands of pixels in graphics. As a result, GPU works well most of the time. Still, you need something more.
Working of Robomorphic computing:
In this computing system, users provide inputs like parameters of robot, limb layout, how joints move, etc. The system then converts all these physical input parameters into the mathematical matrix.
The system then provides the chip-specific hardware calculations to the course of action of the robot, like limb movement or hand movement. As a result, you will get a specific hardware accelerator for a specific task. So, ultimately you end up increasing the efficiency of the robot’s computing needs.
Furthermore, the resulting chip has great efficiency even with a slower clock rate. For example, the chip performs eight times faster than the CPU and eighty-six times faster than the GPU.
Thus, you can manufacture a custom motion planning chip for any robot. So, robots can calculate efficient and safe motions in any work environment without delay.
In conclusion, robomorphic computing opens up a new era in robotics. Thus, it means faster robots with higher reflexes. It also allows robots to think faster that can unlock all-new behaviors.
Image Courtesy: Boston Dynamics