Legged robots

JetHexa

The ROS1 JetHexa hexapod on Jetson Nano, with lidar SLAM and a 3D depth camera in the higher kits.

NVIDIA Jetson
Suggested for university

JetHexa is a hexapod robot built on ROS and powered by an NVIDIA Jetson Nano, running Ubuntu 18.04 and ROS Melodic. HX-35H smart bus servos with 35kg of torque drive its legs, and its vision uses OpenCV, MediaPipe, YOLO model training and TensorRT acceleration. It comes in four kits: the Standard Kit has a monocular camera, the Intermediate Kit adds a SLAMTEC C1 lidar, the Advanced Kit has an EAI G4 lidar and a 3D depth camera, and the Ultimate Kit adds a 6-channel far-field microphone array for voice control. Hiwonder provides schematics, source code, videos and experiments. A separate listing, JetHexa 2027, runs ROS 2.

Zeta42 is an authorised Hiwonder partner in the Middle East.

Configurations

Choose how it ships.

Model: Standard Kit · Intermediate‌ Kit · Advanced Kit · Ultimate Kit

What it does

Built for.

Lidar mapping and navigation
From the Intermediate Kit up, a lidar lets JetHexa build maps, navigate to set points, avoid obstacles and follow a moving target.
3D depth camera
In the Advanced and Ultimate kits, the Astra Pro Plus depth camera gives depth maps, colour images and point clouds. With the RTAB-VSLAM algorithm, JetHexa builds a 3D colour map for navigation, obstacle avoidance and relocalisation.
AI vision
With either camera, JetHexa recognises and tracks colours, follows coloured lines, tracks a selected target with KCF and reads several AprilTags at once. MediaPipe adds body, fingertip and face detection.
Voice control (Ultimate Kit)
The Ultimate Kit's 6-channel far-field microphone array locates the direction of a voice, takes spoken commands, reads ROS messages aloud with text-to-speech and sends JetHexa to a named position on its map.
Group control and formation
One wireless controller can drive several JetHexa robots in step, and in the lidar kits a group can hold a line or triangle formation while moving.
Simulation
JetHexa is built on ROS and works with Gazebo simulation, and an accurate URDF model in RViz shows mapping and navigation results for debugging algorithms.

Where it is used

Applications.

University ROS courses on walking robots
SLAM mapping and navigation projects (Intermediate Kit and up)
3D vision and point cloud work (Advanced and Ultimate Kit)
Multi-robot formation and group control

Videos

See it run.

Specifications

The full sheet.

Weight: 2.5kg
Size (standing, to antenna tips): 426 × 397 × 238 mm
Material: Full-metal hard aluminium alloy bracket (anodised)
Controller: NVIDIA Jetson Nano and ROS expansion board
Expansion board: Built-in IMU, 2-channel PWM, 9-channel serial bus servo interface, GPIO and I2C ports
Display: OLED display showing controller properties and battery voltage
Camera: Monocular camera (Standard and Intermediate Kit); Astra Pro Plus 3D depth camera (Advanced and Ultimate Kit)
Lidar: SLAMTEC C1 (Intermediate Kit); EAI G4 (Advanced and Ultimate Kit)
Microphone: 6-channel far-field circular microphone array (Ultimate Kit)
Battery: 11.1V 3500mAh 5C LiPo battery, removable
Working time: 60min
Operating system: Ubuntu 18.04 LTS + ROS Melodic
Software: PC software + iOS / Android app
Communication: USB / Wi-Fi / Ethernet
Programming language: Python / C / C++ / JavaScript
Storage: 32GB TF card
Servo: HX-35H intelligent serial bus servo, 35kg torque
Control method: Computer / phone / wireless controller
Package size (Advanced Kit): 39 × 36 × 21 cm
Package weight (Advanced Kit): About 3.6kg