RobotPerf, benchmarking interest for robots grows

Source: Acceleration Robotics, February 20, 2023 [1]
RobotPerf subproject grows the ROS 2 Hardware Acceleration Working Group. More meetings and more people participating in discussions around hardware acceleration and robot compute architectures.

PRESS RELEASE: Acceleration Robotics announces the 1st meeting of the Architectures and Processors Working Group (GTAP) of the Spanish Association of the Semiconductor Industry (AESEMI)

Source: Acceleration Robotics, February 21, 2023 [2]
The meeting will take place remotely on February 28 at 6 PM CET . It will also bring together the international robotics community interested in hardware acceleration and advanced computing architectures for robots.

How ChatGPT can control robots

Source: The Robot Report, February 22, 2023 [3]
Microsoft has extended the capabilities of OpenAI's ChatGPT language model to control a robotic arm and aerial drone. To make effective use of ChatGPT for robotics applications, the researchers constructed a pipeline composed of the following steps: defining a high-level robot function library, building a prompt for ChatGPT, identifying the set of allowed high-level functions from the library, and providing feedback to ChatGPT on the quality and safety of the output code.

How Nvidia dominated AI — and plans to keep it that way as generative AI explodes

Source: VentureBeat, February 23, 2023 [4]
Nvidia's GTC conference will feature a fireside chat with Ilya Sutskever, cofounder and chief scientist at OpenAI, who created AlexNet, the pioneering neural network for computer vision that led to the next decade's AI success stories.

The state of FPGA functional verification

Source: Semiwiki, February 15, 2023 [5]
The Wilson Research Group has been compiling an FPGA report every two years since 2018, and the FPGA market is forecasted to grow to $8.1 billion by 2025. Low volume applications benefit from the NRE of FPGA devices, and FPGAs now include processors.

No chips but semiconductor M&A activity is booming

Source: TechNative, February 22, 2023 [6]
Semiconductor chips have been a hot topic due to surging demand and shortages caused by the pandemic, leading to disruption across industries.

Source: The Robot Report, February 16, 2023 [7]
Robot installations hit an all-time high in 2021, bringing in $2.38 billion, and the IFR identified five trends for the industry in 2023.

Tomorrow's semiconductor workforce

Source: Semiconductor Engineering, February 16, 2023 [8]
The American semiconductor industry is facing a shortage of talent to fill critical manufacturing roles, and the SEMI/MITRE Engenuity 2022 Supply Chain Survey Report found that problems in hiring senior-level engineers indicate significant challenges in both retaining and growing talent. To address this, the industry needs to work cooperatively to attract, retain, and train up the diverse workforce that it requires.

Robotic sessions at GTC 2023

Source: NVIDIA, February 21, 2023 [9]
See how AI, simulation, and synthetic data generation accelerate robotics research and deployment


ChatGPT for robotics: design principles and model abilities

Source: Microsoft, February 20, 2023 [10]
This paper presents an experimental study of OpenAI's ChatGPT for robotics applications, exploring its ability to use free-form dialog, parse XML tags, and synthesize code, as well as task-specific prompting functions and closed-loop reasoning. It shows that ChatGPT can be effective at solving several tasks, while allowing users to interact with it primarily via natural language instructions.

2D mapping of mobile robot based on micro-ROS

Source: iopscience, September 25, 2022 [11]
Micro-ROS-based mobile robot system uses micro-XRCE-DDS middleware to communicate with ROS2 in an extremely resource-constrained microcontroller environment.

Hardware acceleration of lane detection algorithm: a GPU versus FPGA comparison

Source: arXiv, December 19, 2022 [12]
The lane detection algorithm is used in autonomous driving and smart vehicle systems and is divided into two steps: edge detection and line detection. This paper compares the performance of the two systems to evaluate the trade-off for latency, power consumption, and utilization.

Previous Hardware Acceleration in Robotics Newsletters

Past ROS 2 Hardware Acceleration Working Group meetings

  1. ↩︎

  2. ↩︎

  3. Crowe, S. (2023, February 22). How ChatGPT can control robots. The Robot Report. ↩︎

  4. Goldman, S. (2023, February 23). How Nvidia dominated AI — and plans to keep it that way as generative AI explodes. VentureBeat. ↩︎

  5. Payne, D. (2023, February 15). The state of FPGA functional verification. Semiwiki. Retrieved February 22, 2023, from ↩︎

  6. Ward, J. (2023, February 22). No chips but semiconductor M&A activity is booming. TechNative. ↩︎

  7. The Robot Report Staff. (2023, February 16). 5 top robotics trends to watch in 2023. The Robot Report. ↩︎

  8. Jammy, R. (2023, February 16). Tomorrow's semiconductor workforce. Semiconductor Engineering. ↩︎

  9. Robotic sessions at GTC 2023. (2023, February 21). NVIDIA. ↩︎

  10. Vemprala, S., Bonatti, R., Bucker, A., & Kapoor, A. (2023, February 20). ChatGPT for robotics: design principles and model abilities. Microsoft. ↩︎

  11. Bin, L., Zhengguang, M., & Yongguo, Z. (2022, September 25). 2D mapping of mobile robot based on micro-ROS. iopscience. ↩︎

  12. Alshemi, M., Saif, S., & Ther, M. (2022, December 19). Hardware acceleration of lane detection algorithm: a GPU versus FPGA comparison. arXiv. ↩︎