The rapid growth of the humanoid robot industry has led to the widespread availability of robots with dexterous hands in the market. However, the development of most practical applications is hindered by the scarcity of trained data, and the case-by-case learning process is highly time-consuming.

To address this challenge, this Inno Wish project aims to develop a comprehensive training kit for enhancing dexterous hand functionality. The training kit will enable users to tailor their training to specific tasks based on individual needs. Additionally, the kit will support users to explore the capabilities and limitations of their hands. The ultimate goal of the kit is to establish modular skill sets that can be initially deployed on other humanoid robots for further advancement.

• Training kit set with vision and tactile sensing devices
• Training software which can integrate with both devices and dexterous hands
• Camera to capture and record the training progress
• Data collectors for the tasks’ information (such as movement, speeds and locations)
• Data analytic for the collected big data
• Own VLA model
• 3 training tasks guaranteed to be achieved (to be proposed in the solution)
• Expandable training tasks database