Interfacing and controlling an exoskeleton

2 views (last 30 days)
Nick
Nick on 7 Sep 2015
Hi guys, I would like to keep this as short as possible so here goes. I'm in a student team where we are designing and developing an Exoskeleton based off our university's existing exoskeleton, The MINDWALKER.
We want to change the electrical systems since the ones they are using is undocumented and has a poor gait algorithm. They are using a backpack PC connected running a ginormous simulink model. That PC is connected to an ethercat coupler with several field busses from Beckhoff. Each joint on the exoskeleton has a dedicated FPGA running some sort of control algorithm for each motor (still unknown what exactly). The FPGA also interfaces with the encoder, driver and an IMU. Each FPGA is daisy chained with E-Bus back up to the EtherCAT coupler.
Since no one understands this undocumented system we decided to do the following:
We want to put a real time target machine on the back of the exoskeleton running all control algorithms. Each joint would then have a board with a fast TI MCU for controlling the motor and relaying all sensor data back to the xPC. This would essentially eliminate the EtherCAT system (i don't know why they used this in the first place, will find out tomorrow).
Does this sound reasonable? The MCU we have in mind is a 2kHz TI MCU with InstaSpin for motor control. We are obviously going to test this system before we actually implement it full scale but before we purchase anything we want to know if this idea is even realistic.

Answers (0)

Categories

Find more on Multicore Processor Targets in Help Center and File Exchange

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!