Gasgoo Munich- on Nov 18 that at the 6th Automotive AI Intelligent Chassis Conference 2026, Qi Jianing, deputy general manager of Chenzhi Technology's Shanghai center, revealed the company is advancing chassis fusion control technology based on deep reinforcement learning. In tests comparing it to traditional model-based ESP control schemes, the AI control solution showed improvements in vehicle dynamic response and stability.

According to Qi, the reinforcement learning control scheme demonstrated advantages in vehicle state convergence, speed, and sideslip angle during double- and single-lane change tests. Notably, average secondary positive and negative peak values for certain dynamic response indicators dropped by roughly 60%, while deceleration loss was also cut by about 60%.
The solution employs an Actor-Critic reinforcement learning architecture, continuously optimizing control strategies through a reward feedback mechanism, while a safety constraint layer ensures control commands remain within the physical limits of the actuators. During deployment, the system outputs control commands via forward neural network computation, reducing the need for online solving and enhancing real-time response capabilities.
Regarding its technical roadmap, Chenzhi Technology is exploring an evolution from multi-sensor perception fusion, motion prediction, and decision planning to end-to-end neural network control. The aim is to directly allocate control signals to actuators like braking, steering, and suspension systems to support the chassis motion control requirements of high-level autonomous driving.
Furthermore, the company has deployed an internal AI software R&D platform covering requirements analysis, task planning, automated modeling, verification, and traceability. Plans are in place to expand this to the entire software development lifecycle. The platform is currently undergoing continuous iteration based on actual user feedback.
Qi also revealed that Chenzhi Technology is researching AI-based sensorless tire pressure monitoring models, with future projects set to include visual road friction coefficient prediction, adaptive driving styles, chassis energy consumption optimization, and multi-system coordinated control.
In terms of product layout, Chenzhi Technology has already covered brake-by-wire, steer-by-wire, and intelligent suspension. The company disclosed that cumulative R&D investment in its wire-control products has reached 690 million yuan.









