Hardness detection based on flexible sensor

The flexible sensor is embedded into a self-made pneumatic software hand through silicone pouring method, and the software hand is controlled to grasp the test object, and the hardness of the object is determined by the change in resistance. The experimental process is as follows: fix the soft hand and robotic arm together, and use a 3D printing bracket to fix the tested object on the optical platform; Control the robotic arm to move the soft hand to a position suitable for grasping; Apply a fixed amount of air pressure (set at 35kPa) to the soft hand to achieve a fixed load effect; Maintain the air pressure for 4 seconds and use the acquisition card to collect the waveform of sensor voltage changes. Divide the collected data into four categories according to the type of object being tested, with a total of 120 sets of data for each category. Among them, 100 sets are used as the training set and 20 sets are used as the test set. Use the CatBoost algorithm to analyze the data. The data is collected using a collection card, using only the CH1 channel, where CH1 represents the voltage signal of the sensor collected in the circuit, in volts.