News

ROS 2 Hardware Acceleration Working Group #13

Source: Acceleration Robotics, November 12, 2022 [1]
Interested in making robots faster. Join the ROS 2 Hardware Acceleration Working Group and learn how! In this meeting we will touch on Updates on the Robotic Processing Unit (robotcore) and the Robotics MCU projects and discussion around REP-2014 including the recent community discussion.
Join folks from the robotics industry and the semiconductors industry to discuss how to make robot compute architectures better.

Robotics markets: up, down or just sideways?

Source: Fierce Electronics, November 11, 2022 [2]
Robots as a category are clearly on an upward trajectory, even if there are not investors for every robot product or customer. Some robotics market segments appear to be doing better than others. Amazon also unveiled its first fully autonomous robot, Proteus, in June.

Nvidia offers new advanced chip for China that meets U.S. export controls

Source: Reuters, November 8, 2022 [3]
Nvidia says new chip meets export control rules aimed at keeping cutting-edge technology out of China's hands. The chip, called the A800, represents the first reported effort by a U.S. company to create processors for China. Nvidia has said the export limitations could cost it hundreds of millions of dollars in revenue.

SiPearl works with AMD on GPU support for Arm HPC chip

Source: The Register, November 15, 2022 [4]
Chip designer SiPearl is working with AMD on software support for exascale supercomputing systems. The aim is to ensure that a software ecosystem is in place ready for such applications. It will work on the interoperability of AMD's ROCm software stack for developing GPU accelerated applications.

Transforming robots help to transfer skills

Source: Scientific American, November , 2022 [5]
A new method helps robots transfer skills between different shapes and sizes. Carnegie Mellon computer scientists created a bridge of simulated robots that slowly shifts in shape. Each intermediate robot practices its task until it reaches a threshold success rate, before the controller code is passed on to the next robot.

In need of precise timing and reliability?

Source: New Electronics, November 15, 2022 [6]
In a vehicle, when pressing the accelerator, the vehicle speeds up almost instantaneously – there is no noticeable delay between pressing the pedal to acceleration.
To apply this example to the topic discussed here - if the vehicle were a system the surroundings, pressing the accelerator and the output, that is the vehicle’s speed - the system would be implementing what’s known as “real-time control.”

Chips in space: Reprogrammable AMD AI SoC cleared for liftoff

Source: The Register, November 15, 2022 [7]
AMD's reprogrammable Versal system-on-chip has completed a Class B qualification. The space chip is set to begin shipping in early 2023. Raytheon plans to use it for "next-generation space processors" for spacecraft and other types of equipment.

Researchers help robots navigate crowded spaces with new visual perception method

Source: University of Toronto News, November 9, 2022 [8]
Researchers from the U of T Institute for Aerospace Studies have developed a system that improves how robots stitch together a set of images taken from a moving camera to build a 3D model of their environments.

Scaling mobile robots for success

Source: Modern Materials Handling, November 10, 2022 [9]
AMR deployments have been gaining ground to address labor availability issues. Fleets often start small in pilots and add robots over time, but how do you get to a better system? Success factors include proper solution scoping, RaaS options, integration platforms, and may involve software to orchestrate multiple AMR fleets.

Choosing the right robot for warehousing, manufacturing

Source: Encoder Manufacturer, August 15, 2022 [10]
Automated guided vehicles (AGVs), autonomous mobile robots (AMRs) and static robots have different needs and abilities. Knowing the difference is critical to understand which choice will be the best fit.

NVIDIA grace Hopper Superchip architecture in-depth

Source: NVIDIA Technical Blog, November 15, 2022 [11]
The NVIDIA Grace Hopper Superchip Architecture is the first true heterogeneous accelerated platform for high-performance computing (HPC) and AI workloads. It accelerates applications with the strengths of both GPUs and CPUs while providing the simplest and most productive distributed heterogeneous programming model to date. Scientists and engineers can focus on solving the world’s most important problems.

Papers

Monitoring ROS2: from requirements to autonomous robots

Source: arXiv, September 28, 2022 [12]
A formal approach to generating monitors for autonomous robots from requirements written in a structured natural language. Ogma generates monitor specifications from the FRET output, which are compiled into hard-real time C99. The goal is to treat the generated ROS2 packages as black boxes and integrate them into larger ROS2 systems with minimal effort.

Energy efficient hardware acceleration of neural networks with power-of-two quantisation

Source: arXiv, September 30, 2022 [13]
Deep neural networks dominate the domain of most vision systems, providing high performance at a cost of increased complexity. In this paper, we show that a hardware neural network accelerator with PoT weights can be at least 1.4 x more energy efficient than the uniform quantisation version.

Previous Hardware Acceleration in Robotics Newsletters

Past ROS 2 Hardware Acceleration Working Group meetings

  1. https://www.linkedin.com/events/6997171083560804352/comments/ ↩︎

  2. Hamblen, M. (2022, November 11). Robotics markets: up, down or just sideways? Fierce Electronics. https://www.fierceelectronics.com/sensors/robotics-markets-down-or-just-sideways ↩︎

  3. Lee, J. L. (2022, November 8). Exclusive: Nvidia offers new advanced chip for China that meets U.S. export controls. Reuters. https://www.reuters.com/technology/exclusive-nvidia-offers-new-advanced-chip-china-that-meets-us-export-controls-2022-11-08/ ↩︎

  4. Robinson, D. (2022, November 15). SiPearl working with AMD on GPU support for arm HPC chip. The Register: Enterprise Technology News and Analysis. https://www.theregister.com/2022/11/15/sipearl_amd_arm_hpc_chip/?utm_source=daily&utm_medium=newsletter&utm_content=article ↩︎

  5. Hutson, M. (2022, November). Transforming robots help to transfer skills. Scientific American. https://www.scientificamerican.com/article/transforming-robots-help-to-transfer-skills/ ↩︎

  6. In need of precise timing and reliability? (2022, November 15). New Electronics. https://www.newelectronics.co.uk/content/features/in-need-of-precise-timing-and-reliability/ ↩︎

  7. Martin, D. (2022, November 15). Reprogrammable AMD versal AI chip cleared for use in space. The Register: Enterprise Technology News and Analysis. https://www.theregister.com/2022/11/15/amd_versal_space_chip/?utm_source=daily&utm_medium=newsletter&utm_content=article ↩︎

  8. Jinje, S. (2022, November 9). Researchers help robots navigate crowded spaces with new visual perception method. University of Toronto News. https://www.utoronto.ca/news/researchers-help-robots-navigate-crowded-spaces-new-visual-perception-method ↩︎

  9. Michel, R. (2022, November 10). Scaling mobile robots for success. Modern Materials Handling. https://www.mmh.com/article/scaling_mobile_robots_for_success ↩︎

  10. Company, E. P. (August 15). Choosing the right robot for warehousing, manufacturing | Article. Encoder Manufacturer | Rotary Incremental & Absolute Encoders Company. https://www.encoder.com/choosing-right-robot-for-warehousing-manufacturing-operations-article?utm_campaign=FY23 Motors&utm_source=DW - Design World&utm_medium=enewsletter&utm_content=Choosing the right robot for warehousing%2C mfg article (CFE) ↩︎

  11. Evans, J., Andersch, M., Sethi, V., Brito, G., & Mehta, V. (2022, November 15). NVIDIA grace Hopper Superchip architecture in-depth. NVIDIA Technical Blog. https://developer.nvidia.com/blog/nvidia-grace-hopper-superchip-architecture-in-depth/ ↩︎

  12. Perez, I., Mavridou, A., Pressburger, T., Will, A., & Martin, P. J. (2022, September 28). Monitoring ROS2: from requirements to autonomous robots. arXiv. https://arxiv.org/pdf/2209.14030.pdf ↩︎

  13. Rus, D. P., & Kryjak, T. (2022, September 30). Energy efficient hardware acceleration of neural networks with power-of-two quantisation. arXiv.org e-Print archive. https://arxiv.org/pdf/2209.15257.pdf ↩︎