ROSpider ROS2 Hexapod Robot with Jetson/Raspberry Pi, Multimodal Large AI Models, ChatGPT, 3D Vision, SLAM Mapping & Navigation

The ROSpider is an advanced, all-terrain ROS 2 hexapod robot designed for cutting-edge research, education, and development in Embodied AI. Powered by Jetson or Raspberry Pi, it features an eighteen-servo bio-mimetic chassis, a 6DOF robotic arm, and a rich array of sensors including LiDAR and a 3D depth camera. It is ideal for university students, robotics educators, and advanced makers looking to explore multimodal AI models, SLAM navigation, and complex manipulation.

Bio-mimetic eighteen-servo hexapod chassis

Built with eighteen 35KG high-voltage bus servos, the ROSpider utilises advanced bio-kinematic gait algorithms for exceptional flexibility. This all-terrain platform dynamically adjusts its posture and gait combinations to navigate narrow spaces and avoid obstacles.

Integrated multimodal Embodied AI

By deploying large vision, voice, and text AI models, the platform processes data from its LiDAR, camera, and microphones. This unlocks advanced Embodied AI applications such as emotion perception, semantic understanding, and AI-driven navigation.

Depth vision and robotic manipulation

The robot fuses real-time RGB and depth data to perceive precise colour and point cloud details. Paired with inverse kinematics, its integrated 6DOF robot arm performs complex 3D tasks including target tracking, grasping, sorting, and transport.

Precision SLAM mapping and navigation

Equipped with TOF LiDAR, encoders, and an IMU, the robot achieves precise SLAM mapping and autonomous navigation. It detects dynamic obstacles in real-time to optimise path planning in complex environments.

Intelligent scene understanding

Utilising a large vision model, the robot analyses its surroundings to assess lighting conditions, identify objects, and determine attributes. It provides immediate feedback on these visual scenes via text or voice.

Real-time visual object tracking

The advanced vision model enables the robot to automatically identify and lock onto specific targets. It maintains stable, real-time tracking even when operating in complex environments.

Semantic voice control

Powered by a large speech model, the integrated 6-microphone array accurately interprets natural voice commands. This allows users to execute complex robotic actions through seamless, conversational interaction.