Pudu Robotics has introduced the PUDU D5 Series, a line of autonomous quadruped robots designed for industrial and outdoor environments. The series consists of the PUDU D5, a legged model intended for uneven and unstructured terrain, and the PUDU D5-W, a wheeled version suited for mixed-surface conditions.
According to the company, the robots are built for tasks in environments that remain difficult to automate, including areas with uneven surfaces, slopes, staircases and variable weather exposure. Pudu Robotics states that the D5 Series is intended for applications such as inspection, patrol, logistics support and emergency response.
In a statement, Founder and CEO Felix Zhang said the product line advances the company’s broader strategy of developing multiple robot types that can be used together across different scenarios.
The robots incorporate a dual-processor computing system based on NVIDIA Orin and RK3588. Pudu Robotics reports that this architecture supports real-time SLAM, 3D reconstruction, object recognition, obstacle avoidance and path planning. The perception system combines fisheye cameras and dual LiDAR sensors to generate 3D point-cloud data, which the company says enables localization and navigation in changing environments.
The D5 Series is described as capable of mapping and navigating facilities of up to one million square meters. The robots have a stated range of 14 kilometers per charge and are designed to manage autonomous workflows, including departure, patrol operations and return-to-charging procedures. The platform supports payloads of up to 30 kilograms, with aluminum construction and torque control intended to maintain stability.
Mobility features include wheel-leg hybrid locomotion, speeds of up to 5 meters per second on flat surfaces, the ability to climb 30-degree slopes and clearance for 25-centimeter steps. The company specifies an operating temperature range from minus 20 to 55 degrees Celsius and an IP67 rating.
Additional functions include autonomous following based on combined vision and depth sensing, as well as gesture and voice interaction supported by a gesture library and a microphone array with noise reduction.
