Debashis Das writes about the D-Robotics RDK S100P, a robotics single-board computer featuring a 6-core Cortex-A78AE CPU, a 4-core Cortex-R52+ real-time MCU domain, and a Nash BPU rated at 128 TOPS INT8, positioned against the NVIDIA Jetson Orin NX 16GB. The kit ships with 24 GB LPDDR5, runs an Ubuntu 22.04-based RDK OS with ROS 2 Humble, and includes expansion boards for additional MIPI/GMSL camera inputs and CAN FD ports.
- The R52+ cores support lockstep (safety) and split-lock modes, potentially eliminating a separate real-time controller
- Priced at $899 (DFRobot) or $999.99 (Waveshare); a $625.99 Amazon listing is the older RDK S100 with only 80 TOPS
- TOPS figures are not directly comparable to NVIDIA's since the two vendors measure inference performance differently
- The MCU expansion board adds 5× CAN FD at 8 Mbps, but its BMI088 IMU was unsupported in the V4.0.3 SDK
- Board supports up to 150 W max power draw in a 121 × 120 × 51 mm enclosure
- The R52+ cores support lockstep (safety) and split-lock modes, potentially eliminating a separate real-time controller
- Priced at $899 (DFRobot) or $999.99 (Waveshare); a $625.99 Amazon listing is the older RDK S100 with only 80 TOPS
- TOPS figures are not directly comparable to NVIDIA's since the two vendors measure inference performance differently
- The MCU expansion board adds 5× CAN FD at 8 Mbps, but its BMI088 IMU was unsupported in the V4.0.3 SDK
- Board supports up to 150 W max power draw in a 121 × 120 × 51 mm enclosure
OLogic provides full-cycle robot engineering services, guiding projects from initial concept sketches through to final manufacturing and market readiness. Their methodology focuses on iterative prototyping to facilitate early testing and design refinement throughout the development lifecycle. The service covers essential stages such as product requirements planning, electrical engineering specialized for motor control and sensor networks, industrial and mechanical design including thermal analysis, software development, compliance testing for FCC/UL/CE standards, and comprehensive design for manufacturing support.
A collection of various LiDAR and ToF (Time-of-Flight) sensing solutions including 360 degree laser range scanners, depth cameras for ROS, and single-point ranging sensors. The selection features technologies such as DTOF, solid-state LiDAR, and RGB-D cameras suitable for SLAM and interactive AI applications.
Key product types:
RPLiDAR series (A1M8, A2M12, S2L, etc.)
STL DTOF sensors
CS depth cameras for ROS
TF series single-point ranging sensors
YDLIDAR linear array solid state LiDAR
Key product types:
RPLiDAR series (A1M8, A2M12, S2L, etc.)
STL DTOF sensors
CS depth cameras for ROS
TF series single-point ranging sensors
YDLIDAR linear array solid state LiDAR
The Metis M.2 card is a high-performance AI inference accelerator designed for constrained, small-footprint devices. Powered by a single quad-core Metis AIPU, it enables state-of-the-art AI capabilities including multi-camera inference and support for multiple independent parallel neural networks. The card offers seamless integration via the Voyager SDK and maintains high prediction accuracy through advanced quantization tools.
NocKinematics is a modern, modular, and lightweight C++ Inverse Kinematics library designed specifically for Arduino and ESP32 microcontrollers. It utilizes the FABRIK (Forward And Backward Reaching Inverse Kinematics) algorithm to provide fast, iterative computations that are more efficient than traditional Jacobian Matrix approaches. The library is optimized for memory-constrained systems like AVR and ESP8266 by using specialized dynamic memory allocation to prevent RAM fragmentation.
Key features and topics:
* N-Joint Support for arbitrary numbers of connected joints.
* Memory-optimized architecture avoiding heavy std::vector usage.
* Platform agnostic compatibility with Arduino Uno, Nano, Mega, ESP8266, and ESP32.
* Practical implementation examples ranging from basic logic verification to multi-DOF servo motor control.
* Support for complex mechanisms like snake or tentacle simulations via the MultiJointSnake example.
Key features and topics:
* N-Joint Support for arbitrary numbers of connected joints.
* Memory-optimized architecture avoiding heavy std::vector usage.
* Platform agnostic compatibility with Arduino Uno, Nano, Mega, ESP8266, and ESP32.
* Practical implementation examples ranging from basic logic verification to multi-DOF servo motor control.
* Support for complex mechanisms like snake or tentacle simulations via the MultiJointSnake example.
An introduction to JD Robotics and the Stewy V2 platform, designed as an approachable learning tool for robotics. The project emphasizes open-source development, simple yet capable hardware designs that maximize motor utility, and active community support for troubleshooting and customization.
This article introduces ROSA, a Robot Operating System (ROS) framework designed to seamlessly integrate Large Language Models (LLMs) into embodied AI systems. ROSA addresses the challenges of connecting LLMs to robotic hardware by providing a standardized interface for perception, planning, and action.
The framework utilizes a prompt-based approach, converting robot tasks into natural language prompts for the LLM. This allows for flexible task specification and reasoning.
ROSA also includes tools for managing LLM outputs, ensuring safe and reliable robot behavior. The authors demonstrate ROSA’s capabilities through various experiments, showcasing its potential for creating more intelligent and adaptable robots.
The framework utilizes a prompt-based approach, converting robot tasks into natural language prompts for the LLM. This allows for flexible task specification and reasoning.
ROSA also includes tools for managing LLM outputs, ensuring safe and reliable robot behavior. The authors demonstrate ROSA’s capabilities through various experiments, showcasing its potential for creating more intelligent and adaptable robots.
Researchers at Huawei, Technical University of Darmstadt, and ETH Zurich have developed a new framework integrating LLMs with the Robot Operating System. This framework aims to improve robots' ability to understand and execute natural language instructions.
The system translates user commands (like "pick up the green block") into actionable steps for the robot, using either direct code execution or a decision-making "behavior tree." Testing has shown promising results, with robots successfully completing tasks using open-source LLMs. This work, published in *Nature Machine Intelligence*, represents a step towards more versatile and responsive robots capable of operating in real-world environments. The code is available as open-source.
The system translates user commands (like "pick up the green block") into actionable steps for the robot, using either direct code execution or a decision-making "behavior tree." Testing has shown promising results, with robots successfully completing tasks using open-source LLMs. This work, published in *Nature Machine Intelligence*, represents a step towards more versatile and responsive robots capable of operating in real-world environments. The code is available as open-source.
Thigmotaxis is a behavioral response to tactile stimuli, referring to an organism's movement in reaction to physical contact with surfaces or objects. It manifests as wall‑following or centrophobic behavior, where animals tend to stay near perimeters in enclosed spaces. The response is genetically grounded and observed across species, including rodents, insects, and fish. Positive thigmotaxis involves moving toward or staying in contact with a surface, often for shelter or safety, while negative thigmotaxis involves moving away from contact. Mechanisms involve mechanoreceptors such as tactile receptors in vertebrates or setae and antennae in invertebrates, which relay sensory input to the nervous system to trigger motor responses. Researchers use thigmotaxis as an assay for anxiety, stress, and navigation in animal models.
Sowbot is an open robotics platform designed to scale regenerative agriculture by providing accessible, lightweight robotics to researchers and farmers. It aims to bridge the "prototype gap" in agricultural robotics with a Reference Hardware Design and a Production-Ready Software Stack. The system comprises an 'Open Core' compute unit with dedicated boards for control/safety and perception/AI, powered by open-source software like Lizard, RoSys, and DevKit ROS. Various platforms like Sowbot, Sowbot Mini, and Sowbot Pico cater to different development stages, all emphasizing modularity, open-hardware, and a commitment to sustainability.
