Towards robot accelerators, democratization of hardware acceleration in robotics


VITORIA-GASTEIZ, Spain–(BUSINESS WIRE)–Acceleration Robotics, a robotics start-up based in the Basque Country, Spain, with Professor Vijay Janapa Reddi of Harvard University and researchers from the Harvard Edge Computing Lab will present their latest work on democratize the use of hardware acceleration in robotics in a scalable, vendor and technology independent way. Entitled “RobotCore: an open architecture for hardware acceleration in ROS 2“, their paper describes and discloses a reference implementation of the architectural pillars and programming conventions needed to introduce hardware acceleration into robotics in a sustainable way, avoiding semiconductor vendor lock-ins. In a short time, the Hardware acceleration enables the creation of custom compute architectures (known as ‘accelerators’ or robot ‘cores’) that leverage computational parallelism.Instead of relying solely on processors, with hardware acceleration (via FPGAs or GPUs), roboticists can empower faster robots, with reduced computation times, lower power consumption and more deterministic responses. The central idea behind their research is to facilitate the process of using this technology which, according to the researchers, with the current solutions requires “expertise in each specific hardware (acceleration) platform”.

The research effort is led by Víctor Mayoral-Vilches, an experienced roboticist and former system architect at Xilinx (now AMD) who left the company to create his own startup (Acceleration Robotics) which focuses on creating these robot accelerators, or semi-conductor bricks for robots combining CPU, FPGA and GPU. The entire group demonstrated together its open architecture against the Robot Operating System (ROS 2), a popular event-driven programming interface for creating robotic applications. The result robot accelerators offer significant computational speedups compared to the performance of modern processors:

Robots are deterministic machines. Timeliness in their calculations is the most important feature. Their behaviors take the form of computational graphs, with data flowing between nodes, through physical networks (communication buses) and when mapping underlying sensors and actuators. The popular choice for building graphics for robots these days is ROS. Most companies that build real robots use ROS or similar event-driven software frameworks, but they do so in a CPU-centric way. We have demonstrated in our work how hardware acceleration properly combining FPGAs and GPUs can revolutionize robotics, increasing determinism and accelerating robot response times enabling new applications with robots.” Victor Mayoral-VilchesAcceleration Robotics, Founder.

This research also allows us to easily benchmark and benchmark different robotics computing solutions and paves the way for roboticists to prototype hardware designs for silicon robots while maintaining common robotics development practices. Their results were demonstrated with ROS 2, a popular robotics development kit, but it can also be used with other robotics frameworks. Their article was accepted at the International Conference on Intelligent Robots and Systems (IROS), one of the most prestigious meetings in robotics, and will be presented next October in Kyoto (Japan).

Additional Resources

About Acceleration Robotics

Acceleration Robotics is a company that focuses on designing custom brains for robots to speed up their response time. Founded by top robotics experts to provide semiconductor building blocks for robots, the company leverages GPUs and FPGAs to create custom hardware that makes a robot run faster.

About Edge Computing Lab

The Edge Computing Lab resides at the John A. Paulson School of Engineering and Applied Sciences at Harvard University. Its team of computer systems architects specialize in advanced computing platforms, with expertise in understanding the interactions between circuitry, architecture, and software layers.


Comments are closed.