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
- Build the dockerfile
sh ./docker_configs/dockerfile.sh | docker build --build-arg APP_UID=$(id -u ${USER}) -t tactile_sensor_velostat -f - .
-
- 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)
- Or
docker-compose -f docker-compose_nvidia.yml up
if you have nvidia on a laptop, ordocker-compose -f docker-compose.yml up
if other (intel, amd) $ sudo xhost +local:root
docker exec -it --privileged tactile-sensor-velostat_master_1 bash
start to work in docker
- Either use
- 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
- TODO