Gasgoo Munich- On September 19, 2026, PITA—the general intelligence model brand under GENISOM AI—released its "Multi-Legged Collaborative Transport" system. The system enables multiple robots to autonomously complete collaborative transport tasks around a single physical objective within unmapped environments, without the need for pre-built maps.
On the operational front, the system shifts human-machine interaction from individual teleoperation to task-level command. Once an operator wearing VR gear designates a target location from a safe zone, a team of ZhiShen L1-W EDU robots takes over: they autonomously perceive the environment, locate themselves and their peers, and share spatial data to plan transport routes. Navigating complex settings—whether facing pedestrians, low obstacles, or stairs—the robots adjust their formation and path in real time based on available space, obstacle distribution, and load status. The system employs an integrated control approach that unifies robots, payloads, and the environment. By leveraging online load estimation, joint trajectory optimization, and whole-body coordination, it ensures synchronized movement and stable physical coupling across the fleet.
Image source: GENISOM AI
The system targets heavy-load scenarios in complex environments where human entry carries high risk and a single robot is insufficient. Specific applications include the coordinated transport of heavy tools and casualty evacuation in earthquake rubble, hazardous material handling in chemical or radiation zones, and the transport of heavy industrial spare parts across intricate factory floors or rugged mountain terrain. PITA, a general intelligence model brand launched by GENISOM AI in 2026 for the physical world, has built a technical framework spanning physical simulation, data generation, model training, and real-machine deployment. Its core portfolio includes general robot models, differentiable physics engines, high-fidelity video world models, and cross-embodiment computing terminals.
