• Introduction
  • Motivation
  • Experimental setup
• Quick start
  • 1. Start to work in computer environment
    • Docker useful commands
  • 3. Update code on STM32
  • 4. Calibrate force sensor
• Useful links

This repository was made for support a scientific paper "Design of a low-cost tactile sensor based on Velostat mounted on curved shapes". The goal of this paper to design a low-cost tactile sensor, which can be used to solve terrain classification problem. It means that we need to mount these sensors on C-like leg.

This repo contains code for:

• Arduino: send data from ground truth 1DOF force sensor Futek LTH350 to PC
• UR5: control manipulator and write down timestamps when it reach needed points
• Velostat tactile sensor: receive data from STM32, decode it and write down to file
• STM32: to receive data from sensor and encode it for sending to PC via RS485
• Postprocessing: Matlab code for making plots

field and UR5e

1. Build the dockerfile sh ./docker_configs/dockerfile.sh | docker build --build-arg APP_UID=$(id -u ${USER}) -t tactile_sensor_velostat -f - .
2. • Either use sh start_to_work.sh (open 3 term tabs, with git, docker-compose and docker exec) (might be needed to do 2.2 before it, if simulator won't work)
   1. Or docker-compose -f docker-compose_nvidia.yml up if you have nvidia on a laptop, or docker-compose -f docker-compose.yml up if other (intel, amd)
   2. $ sudo xhost +local:root
   3. docker exec -it --privileged tactile-sensor-velostat_master_1 bash start to work in docker
3. Start to work

• docker rm -f $(docker ps -a -q) clean all containers from pc
• docker rmi -f $(docker images -a -q) clean all images from pc

1. TODO

• Futek LTH350 datasheet