For WP6, for the tire use-case, we started to implement the developed lattice-based tracking and servoing algorithms on the Michelin use case.
This required additional adaptations to take into account the constraints of the problem including an extra point cloud in the scene and continuous shape of the tire tread.
In addition, we are working along side with Mich in integration of different technique in the robotic cell such as learning from demonstration, treads alignments, Human-Machine-Interface, etc.
Furthermore, for the doll use-case, a 3d reconstruction effect is implemented using the signal from DIGIT tactile sensor.
One multi-layer perception neural network is trained to achieve the better fidelity of 3D contact shape reconstruction of DIGIT tactile sensor.
Based on this developed capability, the contact tension between toy pieces can be observed and adjusted to satisfy the requirements of assembly.
The final experiments related to tactile sensing in doll assembly is being carried out.
Furthermore for WP8, CA-INP is invited as a speaker in ERF2023 conference to present SoftManBot project.