The uOttawa Masters Engineering group made progress on 3D printing the pan-tilt mechanism for the IR camera. Beck gave a tutorial to the group on how to 3D print with the Tinkerines and how to slice the models. Meanwhile, another member of their group worked on installing Tensorflow on the Raspberry Pi. The servos fit perfectly into the mount, and it gets assembled with M2 screws. Beck and Brenda made progress on their Bowie Brain Kits as well. Brenda had a thought about a 3D object detection method, so we brainstormed a bit about that on the whiteboard. Next steps is to continue the work.
Here’s part 2 of the beta brain kit progress! Tonight Beck and Brenda were soldering many of the headers needed for this board, taking approximately 3 hours. Watch the video snippet update here
Some of the bugs found this week were:
Might want to mention about having a micro USB cable that can do sync
- Part 5, Step 16 – Rephrase 4x 5 pins, so they can all be cut at once. Had to sand F headers to fit
- Part 5, Step 25 – Two right-most holes for the current sensor
- Part 5, Step 26 – ‘All available pads’, not quite
They are making great progress and finding useful bugs to improve on in the instructions. I think that next week they will be pretty close to getting set with installing the Teensy and Arduino software to program the board with its first blink! Want to see part 1 of the progress? Check it out here
We met with some of the Master’s engineering design group again. Tonight, they performed benchmark analysis on Bowie the robot. We also discussed in depth the tradeoff predicament version 1.0 of the robot is at with regards to the power of the motors. It was great that they were attentive and understood. We watched some of the videos of the testing of the robot in Summer 2018 to analyse some of the timing of the turns and going forward. At the end, we went through the photos of all 23 iterations of the robot. This was the last piece of project management work – next week is design time!
Today was a big day in terms of Robot Missions milestones! Beck and Brenda are the first beta testers of assembling the Bowie Brain Kit! Woohoo! They will be following along all the steps to get their own Bowie brain alive. This was a pretty epic work session. They spent 5 hours working away on it, by the end, they didn’t even realise how much time had passed! It was 11:30pm by the time we left. 😀 Here is a video update:
Going from a small amount of previous soldering and robot experience, to now soldering this board, they mentioned that the instructions and photos have been super helpful. It was amazing to see their progress on this. Sometimes there were mistakes, but it was ok because we fixed them along the way.
Beck and Brenda found a lot of bugs that should be fixed! Here’s the list of them:
Just remember – if something is extending over the edge of the board before soldering, it’s probably in the wrong place. Avoid cold solder. Resistors do not have directionality, but most of the other parts do. Oh yeah, and you can remove the tape after it’s done being used to hold the soldered component in place.
- Part 2, Step 13 – Should be resistors inserted into R10 and R11 (it’s plural). Make sure the leads of both resistors are not touching.
- Part 2, Step 16 & 17 – It is not clear what direction the diode should go in. Polarity
- Part 2, Step 18 – ‘Resistor’ should be ‘diode’ 😛
- Part 3, Step 3 – Need to write about using tape instead of twisting the leads, because the leads from the speaker can snap and break
- Part 3, Step 8 – ‘Resistor’ should be ‘LED’ 😛
- Part 3, Step 12 – LEDs not resistors
- Part 4, Step 10 – Could also add something about using tape to secure the capacitor in place prior to soldering it and snipping the leads
- Part 4, Step 2 & 3 – There is no directionality for the PN2222A, should mention what side the flat gets inserted
- Part 4, Step 6 – Directionality?
- Part 4, Step 8 – Do not flip the board
- Part 4, Step 14 – Needs clarification that the row of holes the screw terminals go in are the row of Z3,2,1 – not the holes nearest the edge of the board.
- Part 5, Step 1 – Should be Xbee headers (shallow F headers) instead of just shallow F headers to correspond with the packing list properly
- Part 5, Step 2 – Add a bit about keeping the headers perpendicular to the board, not inserting them crooked
- Part 5, Step 4 – Should also mention about how adding in an Xbee as a placeholder will improve the reliability of these headers (just based on previous experience of the forks getting jammed)
- Part 5, Step 5 – You don’t have to flip the board
- Part 5, Step 8 – Orientation
Way to go Brenda and Beck on this progress! More to come next week.
We met with the University of Ottawa Master’s Engineering Design project team tonight, where we showed them a tour of the Prototyping Lab at Bayview Yards and discussed their project plan. We tried to determine what resources we might need for achieving all aspects of the project plan.
The group is super curious about the robot, and robotics in general. Their module will be useful, to see if IR light is useful in detecting plastics. It will be interesting to try it both in the day, and at night.
We received info from the Master’s engineering design team about their project plan. Their project plan is very focused, which is good to hopefully provide a useful result. As well, it contains a stretch goal. Later this week, we will have a chance to meet and discuss what resources will be needed.
Here is some info from their plan:
Our project would be divided into 3 parts, 2 fundamental parts and 1 advanced part
1. Mechanical design of the mechanical structure (Pan-Tilt Mount) – Fundamental
Object: This pan-tilt mount should be able support 360 version of camera
2. Environment configuration of robot vision system – Fundamental
Object: Installing tensorflow on Raspberry Pi, connecting Raspberry Pi and IR camera. After
configuration, the picture captured by IR camera should be able displayed in tensorflow on Raspberry Pi
3. Object detection and recognition based on IR image – Advanced
Object: Designing object detection and recognition algorithm based on IR image using Tensorflow
We look forward to working with them and for developing a new Bowie module. More to come later in the week!