omni-bot update

Its time to give the omni_bot some love changes are as follows

• change rpi4 for a I5 mini pc
• use a t265 for localisation
• hobby servos for steering
• 
  

the omni_bot drives well the wheel control  has to use velocity commands not PWM.

the T265 uses a non standard TF tree and has trouble starting . i will change it for tracking wheel to provide odometery

It moves

Completed the last of the wiring so it was time to give the omni_bot a run the good news is it didn't tear itself apart all movement is in the correct sense. the drive motors don't have encoders at the moment so we just using PWM to control the speed. 30% PWM gives a good speed so I can use smaller motors that are cheaper. The joint state message error I had in the previous post was just a missing header in the message.

Problems

• no reverse
• need a way to reset motor controllers
• insert a delay to allow the steering actuators complete moving before drive starts

Taking shape

The Omni_bot is taking shape the new struts are working the magnetic encoders are calibrated. at the moment we are using ROS Kinetic until the basics are working. We are incorporating the lessons learnt with Robbie. the on-off is only a push button and power to the motors is controlled by the Arduino. and the charger has a spring to give a positive connection. Power is a 2 stage design a hot Bus that handles charging and starting the system and MOSFETs to toggle other systems as required

Pressing problems

• Jointstates not publishing in rviz
• steering turns continuously on shut down

Progress

The Omni_bot is taking shape we have printed the drive struts they seem just as strong as the old struts

Next, task is to add the controller boards and power systems

The return of the omni_bot

Back when I first started building Robbie in 2011 he had a holonomic base with powered castors. In July 2013 I changed Robbie to a differential drive base and incorporated the inmoov torso. After operating Robbie for a while I found the differential drive to be lacking. In the tight environment of a home full of furnisher, he would never be able to function as desired. I decided to relook at the old Omni drive base back in 2011 I made choices that were limited to my experience and tooling. After reviewing my files I found the main cause of the problems encoders were not reliable, this proved to be an ongoing issue with Robbie until they were replaced. The other was inexperience in just about everything to with robotics I feel now is the time to try again. Some parts will be reused others will be redesigned. The big changes are Raspberry Pi 3 in place of the I3, adruino due in place of the mega, better encoders and motor drivers. But the biggest change will be running ROS2. ROS2 is running on the RPI and communicates with the server. Now let the fun begin.