Currently, the industry's typical dexterous hand drive and control solutions primarily use MCUs as the core to control motor drivers, paired with coreless motors and ball screw transmission mechanisms. These can coordinate multiple joints to complete actions like grasping. However, this approach has obvious limitations. Traditional MCUs struggle to handle complex algorithms, making it difficult to achieve the required precision in torque and speed regulation when driving high-precision motors. Furthermore, a single MCU can optimally drive only 3-4 motor channels, making it difficult to meet demands of 12 or more channels. Forcibly increasing the number easily leads to latency.
We employ a T40-based product featuring a heterogeneous computing architecture and a customized logic control scheme. Core motor control algorithms are implemented via hardware logic, reducing software-level latency and effectively ensuring motion control precision, making it suitable for fine manipulation scenarios. It integrates up to 12 independent motor drive channels, supporting the synchronous movement of multiple joints in a dexterous hand and meeting the demands for complex action execution. Through deep coordination and optimized scheduling between "programmable logic units and embedded processing cores," it achieves a balance of high real-time performance and flexible functional adaptation, delivering superior performance in terms of response efficiency and power consumption balance.