The ROS 2 Hardware Acceleration Working Group (HAWG) launched new projects over 2022 including RobotPerf, the Robotics MCU or the Robotics Processing Unit and achieved a 3.7x growth in outreach and over 2021, engaging with more than 1M roboticists within the robotics WG activities during the last year.

The ROS 2 Hardware Acceleration Working Group (HAWG) launched new projects in 2022 including RobotPerf, the Robotics MCU or the Robotics Processing Unit and achieved a 3.7x growth in outreach over 2021, engaging with more than 1M roboticists within the robotics WG activities during the last year.

The ROS 2 Hardware Acceleration Working Group (or HAWG for short) is an open and community-driven robotics group that drives the creation, maintenance and testing of hardware acceleration kernels on top of open standards for optimized ROS 2 and Gazebo interactions over different compute substrates, including FPGAs and GPUs.
This dissemination report is a comprehensive review that presents a summary of the impact and key projects developed by the Working Group over 2022. This working group, consisting of experts in robotics and computer architecture (hardware acceleration), and was formed with the goal of exploring ways to improve the performance and efficiency of robotic systems using hardware acceleration techniques.

The report herein follows from the one produced in 2021 and is intended to serve as a resource for those interested in learning about the latest developments in hardware acceleration for robotics and to provide guidance for future research and development in this area.

New projects during 2022

  • RobotPerf, RobotPerf provides an open reference benchmarking suite that is used to evaluate robotics computing performance fairly with ROS 2 as its common baseline, so that robotic architects can make informed decisions about the hardware and software components of their robotic systems.
  • Robotics MCU: an open source software and hardware ROS 2 microcontroller unit (MCU) powered by RISC-V. The project's goal is to design and develop an open source software and hardware robotics microcontroller unit (Robotics MCU) powered by RISC-V that delivers lower latency and additional real-time capabilities in ROS 2 MCU interactions. The ultimate objective of the project is to design an MCU that contains a native ROS 2 hardware implementation.
  • Robotics Processing Unit: a robot-specific processing unit that uses hardware acceleration and maps robotics computations efficiently to its CPUs, FPGAs and GPUs to obtain best performance. In particular, it specializes in improving the Robot Operating System (ROS 2) related robot computational graphs on underlying compute resources.

Tracking progress

Key milestones

Year Objective
21 ✅ 1) Design tools and conventions to seamlessly integrate acceleration kernels and related embedded binaries into the ROS 2 computational graphs leveraging its existing build system (ament_acceleration extensions) [1], meta build tools (colcon-acceleration extension) and a new firmware layer (acceleration_firmware) [2].
21 ✅ 2) Provide reference examples and blueprints for acceleration architectures used in ROS 2 and Gazebo.
21 ✅ 3) REP-2008 (ROS 2 Hardware Acceleration Architecture and Conventions) first draft proposed PR
22 ✅ 4) Facilitate testing environments that allow to benchmark accelerators with special focus on power consumption and time spent on computations (see HAWG benchmarking approach, community#9, tracetools_acceleration, ros2_kria)
22 ✅ 5) Survey the community interests on acceleration for ROS 2 and Gazebo (see discourse announcement, survey, result).
22 ✅ 6) Produce demonstrators with robot components, real robots and fleets that include acceleration to meet their targets (see acceleration_examples).
22 ✅ 7) Acceleration of complete ROS 2 computational graphs
22 [3] 8) Merge first hardware accelerators (kernels) into upstream packages (candidate: image_pipeline, see image_pipeline instrumented at #717 )
22 ✅ 9) Documentation and a "methodology to hardware accelerate a ROS 2 package" (see #20)
22 [4] 10) Organize workshops on robotics and ROS 2 Hardware Acceleration
22 🚧 11) Robotic Processing Unit, first demonstrators #25
22 ✅ 12) REP-2008 (ROS 2 Hardware Acceleration Architecture and Conventions) accepted as an official standard
22 ✅ 13) REP-2014 (Benchmarking performance in ROS 2) first draft proposed PR
23 🚧 11) REP-2014 (Benchmarking performance in ROS 2)
23 🚧 14) More vendor support. Will work to enable new silicon solutions
23 🚧 15) Robotics MCU #31 milestone 1
23 🚧 16) RobotPerf benchmarks
23 🚧 17) Robotic Processing Unit, first demonstrators #25

Key articles and posts in traceable digital media

Publication (date & link) views reactions
LinkedIn - Reconfigurable computing for robots, an open source approach (surprisingly similar 😂 to this other paper 😆) 10K views 67 reactions
LinkedIn - Join the #hardware #acceleration (re)evolution in #robotics 5K views 41 reactions
LinkedIn - Learn "how #ROS 2 is easing #hardware #acceleration for #robotics" with an open and vendor-agnostic architecture 27K views 295 reactions
LinkedIn - ReconROS Executor: Event-Driven Programming of FPGA-accelerated ROS 2 Applications 10K views 99 reactions
LinkedIn - HAWG#6 - A methodology for ROS 2 #hardware #acceleration that allows speeding up #ROS computational graph dataflows a 20% up 7K views 96 reactions
LinkedIn - Call for HAWG#7 4K views 40 reactions
LinkedIn - PiCAS, a #ROS2 Priority-Driven Chain-Aware Scheduling 15K views 130 reactions
LinkedIn - We compare #FPGA and #GPU performance in ROS 2 perception graphs and propose the Robotic Processing Unit (#RPU), robot-specific #processors that aim to obtain best #ROS peformance. 32K views 400 reactions
LinkedIn - Meet REP-2008, ROS 2 Hardware Acceleration Architecture and Conventions 8K views 75 reactions
LinkedIn - #Robots moving faster require faster #perception computations. We show how #ROS 2 Nodes in AMD's FPGAs are up to #500x faster than in NVIDIA's GPUs 12K views 107 reactions
LinkedIn - RobotCore: An Open Architecture for Hardware Acceleration in ROS 2 25K views 263 reactions
LinkedIn - Message Flow Analysis with Complex Casual Links for Distributed ROS 2 Systems 7K views 61 reactions
LinkedIn - #Yocto and #ROS 2 allow you to build a #production-grade custom operating system for #robots 9K views 93 reactions
LinkedIn - FogROS 2: An Adaptive and Extensible Platform for Cloud and Fog Robotics Using ROS 2 17K views 180 reactions
LinkedIn - Nominated in the #ROS #Awards 2022 as best developer 7K views 82 reactions
LinkedIn - Robotic Computing on FPGAs: Current Progress Research Challenges, and Opportunities 15K views 152 reactions
LinkedIn - Is Modular Architecture Enough? 8K views 51 reactions
LinkedIn - ROBOTCORE to speed-up ROS robots with hardware acceleration 6K views 71 reactions
LinkedIn - The #ROS 2 #Hardware #Acceleration #Stack 11K views 129 reactions
LinkedIn - #ROS 2 Type Negotiation (REP 2009) 6K views 76 reactions
LinkedIn - Acceleration Robotics anuncia #ROBOTCORE, un framework para crear #chips robóticos 15K views 156 reactions
LinkedIn - HAWG#6 recording 5K views 55 reactions
LinkedIn - Acceleration Robotics Launch ROBOTCORE™ to Speed-up ROS 2 Robots 58K views 59 reactions
LinkedIn - #Robots are #deterministic machines 10K views 110 reactions
LinkedIn - Acceleration Robotics es una firma de #semiconductores para la #robótica 6K views 50 reactions
LinkedIn - ROS Robotics companies, a list of for profit robotics companies using ROS 45K views 487 reactions
LinkedIn - Acceleration Robotics is collaborating with AMD to design next-generation robotic compute architectures with #ROS 15K views 184 reactions
LinkedIn - Acceleration Robotics and PlanV are collaborating to create a #robotics #microcontroller unit (MCU) powered by RISC-V and ROS 2 15K views 253 reactions
LinkedIn - ROBOTCORE® is a #processing #unit for the #robotics architect 10K views 115 reactions
LinkedIn - ROBOTCORE®, así es el 'cerebro' robótico español diseñado por Acceleration Robotics 7K views 72 reactions
LinkedIn - The #computer #architecture of Tesla's robot brain underwhelms 12K views 82 reactions
LinkedIn - ROS 2 Hardware Acceleration Working Group #11 14K views 91 reactions
LinkedIn - Acceleration Robotics announce and open sources ROBOTCORE® 11K views 232 reactions
LinkedIn - Acceleration Robotics, AMD partner to design robotic compute architectures with ROS - The Robot Report 5K views 120 reactions
LinkedIn - Vijay joins as an advisor Acceleration Robotics 10K views 105 reactions
LinkedIn - Join me at #ROSCon2022 in #Japan to speak about hardware acceleration in robotics and #systems #architecture 6K views 110 reactions
LinkedIn - #Hardware #acceleration can revolutionize #robotics 5K views 45 reactions
LinkedIn - speaking at #ROSCon2022 about #hardware #acceleration and the #ROBOTCORE Framework 8K views 156 reactions
LinkedIn - An open architecture for Hardware Acceleration in robotics and ROS 2 11K views 140 reactions
LinkedIn - My company Acceleration Robotics is acquiring Technoyantra to serve hardware acceleration in Asia 6K views 70 reactions
LinkedIn - ROS 2 Hardware Acceleration Working Group #12 82K views 292 reactions
LinkedIn - Tesla's Optimus brain hardware and software architecture 7K views 142 reactions
LinkedIn - El cerebro robótico 'made in Spain' cuatro veces más potente que el piloto automático de Tesla 42K views 350 reactions
LinkedIn - final vote on REP 2008 is happening in two weeks 8K views 93 reactions
LinkedIn - ROS 2 Hardware Acceleration Working Group #13 112K views 384 reactions
LinkedIn - Acceleration Robotics expands to India and takes over TechnoYantra to grow in Asia 38K views 153 reactions
LinkedIn - Meet #RobotPerf, the benchmarking suite to evaluate #robotics #computing #performance 12K views 140 reactions
LinkedIn - Is Open Robotics acquisition by Intrinsic a good thing? Yes, I believe it is 22K views 259 reactions
Other digital media
Reddit - Multiple posts over the year 311K views N/A
TOTALS 2021 (roughly) 1.1M+ VIEWS (4x growth 2021) 7490+ REACTIONS (3x growth 2021)

Meeting recordings (YouTube channel)

Publication (date & link) views reactions
HAWG#1 887 views (0.55x growth) 24 reactions
HAWG#2 445 views (0.72x growth) 15 reactions
HAWG#3 399 views (0.32x growth) 15 reactions
HAWG#4 171 views (0.94x growth) 8 reactions
HAWG#5 158 views 3 reactions
HAWG#6 348 views 15 reactions
HAWG#7 217 views 2 reactions
HAWG#8 194 views 9 reactions
HAWG#9 335 views 6 reactions
HAWG#10 339 views 5 reactions
HAWG#11 202 views 2 reactions
HAWG#12 172 views 4 reactions
HAWG#13 131 views 1 reaction

Goals for 2023

The Hardware Acceleration Working Group has grown significantly over 2022. The working group has several key objectives for 2023. The first objective is to develop and publish a comprehensive set of benchmarking tools packed in a suite (RobotPerf) for measuring the performance and efficiency of various software components (ROS packages) and hardware solutions commonly used in robotics. This will enable members of the working group to accurately compare and evaluate the performance of different hardware options and make informed decisions about which components to use in their own projects.

The second objective is to increase the number of vendors participating in the working group. By engaging in targeted outreach and engagement efforts, the group aims to bring more vendors on board, which will provide more options for members and increase the overall diversity of hardware components available for testing and evaluation.

The third objective is to increase the number of collaborations with industry partners. This will gather practical experience and feedback on the performance of hardware acceleration techniques in real-world robotics applications and will also increase the working group's visibility and impact.

The fourth objective is to continue investing resources in the ongoing strategic projects announced during 2022. In particular RobotPerf, Robotics MCU and the Robotics Processing Unit.

Altogether, the objectives for the coming year look as follows:

Year Objective
23 🚧 11) REP-2014 (Benchmarking performance in ROS 2)
23 🚧 14) More vendor support. Will work to enable new silicon solutions
23 🚧 15) Robotics MCU #31 milestone 1
23 🚧 16) RobotPerf benchmarks
23 🚧 17) Robotic Processing Unit, first demonstrators #25

  1. See ament_vitis ↩︎

  2. See acceleration_firmware_kv260 for an exemplary vendor extension of the acceleration_firmware package ↩︎

  3. Did not get buy-in from maintainers. Stopped the effort. ↩︎

  4. Discarded for ROSCon and IROS. Lots of work. ↩︎