Author: Eric Shi

Progress Updates Part 4

Following up on the goals for reading week (Feb 20 – Feb 26):

Major Goals for Reading Week:

  1. Complete machining of robot lift, robot enclosure, and supporting brackets to assemble by the February 27th
  2. Complete integration of vision system with robot
  3. Implementation of LIDAR safety systems and real world testing
  4. Tuning of drive PID controller with load on the wheels
  5. Integration and final wiring on February 27th

The project has fallen behind schedule mainly due to the machining of the robot lift, enclosure, and supporting brackets taking much longer than expected. After 11 days of an average of 8 hours of machining, the lift is done pending final tests, the robot enclosure has only a few more holes to drill, and the supporting brackets are complete. However, this heavy delay has blocked electrical work requiring the assembled robot to complete wiring.

As the team lead and project manager, I am finding it hard to keep the project on track given these delays and many courses having big deliverables around the same time. For the first time in the project, the weekly Sunday work session reserved for group work and integration is being canceled due to the busy schedule of group members and the mechanical work blocking further progress.

There are 10 more working days (not including Sunday). The new deadline for mechanical assembly has been extended to end of the day on Monday, March 6th. This is the third extension and probably the last one we can afford to do while making the symposium deadline.
Here are some updates of the mechanical components:

Upper Enclosure Progress Update:

Robot Lift Update:

Sensor Post Update with Temporary Platform: 

 

 

Progress Updates Part 3

This week was our reading week which meant that all classes were canceled and the group got to dedicate a lot more time towards the project. The team set a few major goals for reading week as follows:

Major Goals for Reading Week:

  1. Complete machining of robot lift, robot enclosure, and supporting brackets to assemble by the February 27th
  2. Complete integration of vision system with robot
  3. Implementation of LIDAR safety systems and real world testing
  4. Tuning of drive PID controller with load on the wheels
  5. Integration and final wiring on February 27th

Since the individual components are more complicated, the detailed overview of each part will have their own blog posts coming up. In terms of the overall progress for each of the goals.

Condensed Updates:

  1. Machining took longer than expected, even with 3 group members working on machining actively during reading week, machining of the lift, chassis, and supporting brackets have not yet been completed but it is scheduled to be completed by March 1st.
  2. Vision system integration was successful with robot arm, vision system can identify a point and the robot arm motion planner creates a path to the point. Further calibration is required but the two systems are able to communicate using ROS.
  3. Navigation LIDAR safety algorithms were implemented using simulations, real world testing has begun but is not complete
  4. PID drive system has been tuned with load, will need to tune again once the mechanical assembly is fully complete
  5. Integration is behind schedule and has not started as of Feb 27th

On the Radar:

There are three weeks left for the project, I will be watching the machining closely, the parts need to be finished by March 1st.

Project Centaur Receives $1,000 from Engineer of the Future Fund

From the Engineer of the Future Trust Webpage:

“The Engineer of the Future Trust was launched by Terry Cunningham, in June 2013 on the 30th Anniversary of his graduation with a BASc (Mech, ’83) in Menlo Park, California. The President of EVault™, one of the world’s premier on-premise software and offsite cloud services for data protection and a proud ambassador for Waterloo Engineering in Silicon Valley, Terry kickstarted the Engineer of the Future Trust with a $100,000 gift and a challenge to alumni to match his donation.

With the support of our alumni in Canada and abroad, the Dean of Engineering has created a first in Canada – the Engineer of the Future Trust – a pool of discretionary micro-seed funding for budding entrepreneurs at Waterloo Engineering.”

Project Centaur has received $1,000 of funding from the Engineer of the Future Trust Fund. The team would like to thank the trust fund committee for our reviewing our application and ultimately finding interest in supporting our goals.  We will continue to work away on our first prototype and the support will go a long way.

 

Progress Updates Part 2

This week, the biggest task on my radar is the robot lift. The robot lift is the component that will extend the robot arm’s height by up to 60 cm. This will increase the size of Centaur’s workspace and enable it to reach more inspection points. The robot lift design has been completed and reviewed.

The robot lift utilizes a linear actuator with feedback control to raise the robotic arm during an inspection. The construction of the robot lift is required to be completed by February 25th, updates will be posted along the way.

Adam is currently working away at developing software for the vision system. A standalone test setup was created to verify the performance of the vision system under the following varied conditions:

  • Different lighting
  • Camera angles
  • Background objects

Adam also developed a GUI to control the pan-tilt and view Xtion feed together.

Wesley is working away at fine tuning and debugging the robotic arm motion planning and control software. Software-enforced joint angle limits are being implemented to constrain the workspace of the robot arm.

I have been working on cleaning up the wiring in the motor compartment of the robot arm and implementing our charger and voltage monitors. I’ve also been looking into interfacing with our CAN bus controller from Linux.

Rahul developed a ROS node capable for receiving multiple LIDAR data channels and flagging when the close proximity of the sensors has been breached by an obstacle for the safety system.

PID Controls, CAN Bus, and Everything Motors

Centaur is powered using four DC motors with the following specifications

  • HP: 1/4
  • Volts: 24
  • Gear Ratio: 11:1
  • 86.9 in-lbs torque at 168 rpm

In addition the internal gearing of the DC motors, a belt mechanism is used to further gear down the output shaft to increase torque.

The main purpose of the mobile platform is to position the robot arm so that it can reach the pipes for inspection. Speed control of DC motors can be achieved using simple armature voltage control in an open-loop configuration. However, open loop speed control would make it difficult to position the robot. The robot would also not be able to reject disturbances in open loop control such as changes in floor condition. A feedforward PID controller was implemented using optical encoders as feedback.

The motor controllers used is the Talon SRX by Cross the Road Electronics. The Talons can be driven using PWM or CAN Bus. With PWM, the motor controllers work in open-loop. However, using CAN Bus, it is possible to connect encoders directly to the motor controller and implement a PID controller directly by programming the TALONs. The benefit of this is motor controls is offloaded from the main controller.

Wiring up the CAN bus was straight forward using a 120 ohm terminating resistor. The only trick was to flash the firmware on all of the Talon’s since the default firmware is for First Robotics Competition and has a token that has to be asserted every 10 seconds.

Further tuning of the PID parameters will be done later this week.

Progress Updates Part I

The major milestone for the project is the Mechatronics design symposium on March 17th, 2017. By then it is expected that a functional prototype of Centaur is ready for public demonstration.

In the past few weeks, the team has been busy working towards a few major milestones.

  • February 20th, Robot Arm and Vision System Integration
  • Feb 22nd, Drive system calibration with full load
  • March 1st, System integration and Testing

The goal is by March 1st, all design and construction work has been completed and the remaining time will be spent on integration and writing the software suite.

Here is a breakdown on what each team member has been working on:

 

  • I (Eric Shi) have been working on implementing PID controls for our drive system, remote driving, setting up the onboard computer, implementing our power distribution system, wiring everything up in the drive compartment of the robot.
  • Wesley Fisher has been working on developing the motion planning and simulation package for the robotic arm. He also been working on interfacing with the pipe thickness ultrasonic transducer.
  • Adam Heins has been working on the vision system that will be used to guide the robot arm. Adam is working closely with Wesley in preperation of integrating the motion planning with the vision system.
  • Jesse Lopes has been working on the design and fabrication of the lift mechanism for the robot arm. He is finalizing the simulations and will be machining the lift in the next week
  • Rahul Rawat is simulating the algorithms that be used to help the robot navigate in a known environment using laser scanners