Pi Wars 2024 Challenge 1 – Lava Palava

Pi Wars 2024 Challenge 1 – Lava Palava

In this post we will focus on the first Pi Wars Challenge – Lava Palava.

We will first have a look at the rules as set out by the Pi Wars organisers, and then write our specific thoughts, actions, trials and errors underneath in date order. We hope to get our robot up and ready in time for this first challenge!

Rules from Pi Wars Website:

Aim of the challenge

To drive from one end of the course to the other as fast as possible, outrunning the lava flows from a volcanic eruption.

Control method

Autonomous.

Time limit

5 minutes.

Rules

• The course is a shallow trough of painted hardboard approximately 7m long with walls approximately 70mm high and set approximately 550mm apart.
• The course has been sketched out below. The angles of the turns may be slightly different (although this course is finished already and they’re pretty accurate to the sketch).
  
• The course will be approximately the same width throughout.
• There will be a white line down the middle of the track to aid navigation. The line will be approximately 19mm wide.
• The course will be laser-timed for accuracy. If you are using laser-guided sensors, we recommend building in an ‘ignore delay’ at the start of the run so the sensor(s) don’t get confused.
• The competing robot must cross the finish line for the time to be recorded.
• The robot must be autonomous.  A “start” and a “stop” button are permitted (whether on the robot or on the controller).
• You will need to manually retrieve your robot from the end of the course and return it to the start between runs.
• New: there will be at least one ‘speed hump’ on the course.

Scoring

• Competitors will be ranked according to the total amount of time taken to drive the course three times; the robot with the shortest combined time will take first place.
• Points will be awarded to the top ranked robots according to the Formula Scoring System.

Additional Points

• 225 points will be awarded for each completed run.
• 100 additional points will be awarded for each run where the robot does not touch the sides of the course.
• 275 additional points will be awarded to the robot with the fastest individual run.

Penalties

• Each touch of the walls incurs a 10 second penalty up to a maximum of 30 seconds.
• Robots scraping along the wall(s) will be penalised to the maximum of 30 seconds.
• It is permitted to rescue the robot and place it back on the course at the place where things went wrong once per run without penalty, but the clock will not be stopped.
• An additional rescue is permitted, incurring a 15 second penalty.
• A third rescue is not permitted; instead the run must be abandoned.
• Abandoned or non-completed runs will be penalised by 60 seconds.
• Robots failing to complete all three runs will not be ranked but will accumulate additional points as above.

Initial thoughts

Key is speed and accuracy. Knowing the placement of barrels up front means can gauge route beforehand (serving purpose of) avoiding knocking any barrels over. But of course, if pushing nominated barrel to appropriate zone will need to think carefully beforehand about what order to go in. In terms of control, Wrecks mustn’t thrust too fast forward or backwards. No jerky movements, must be smooth. What can be attached to front to push (some sort of widened scooper) but not too wide, or even better something that can be controlled (possibly by servo motors) to open and close. That would be good. What do you use to recognise colours – is it a camera sensor? Need to explore this. How do you set the parameters?

Brainstorming session – 31 October 2023

So, automatic – 5 minutes and must get across the finish line, and as fast as possible. It must be able to turn.

Initial thoughts

Line following. Set parameters so ‘eyes’ understand width of central line.

Height – 70mm. What is relevance of this?

3 goes. Tweak between? Allowed/disallowed? Formula Scoring System (what is that?). Imperative to complete run (for bonus points). Ensure no touching sides/scraping – Max Speed testing

Discussion Flash – Lucodey and SavLeode–

Autonomous robot – A robot you don’t have to control and Start and stop button – Useful 

We have considered the length of the course. And in terms of being able to turn we need to find out what type of sensor follows a line. We think an ultrasonic sensor (eyes) will help with knowing where the edges are but we think Wrecks will definitely need some sort of sensor that follows a line.

• Make sure ‘eyes’ which are sensor is positioned low on robot. If too high then sensor may not work and need to set parameters so that it stops as certain height.
• Use a sensor (underneath) to follow line. Will need to set parameters so and so mm and if deviates by so and so mm then either stop or reposition. What mechanism can be used if sensor goes beyond width of line (either side) in worst case and veers off. Important as unable to retrieve Wrecks.
• Width 550mm. We believe sensors will have to be able to make the robot right if it veers off from central line. As such, we believe we will need to ensure parameters are set for this.
• Size of Wrecks – Centre of gravity – low for speed.
• Bends – How will line following sensor adjust and how quickly? Or is better to use some type of measurer (to first point in bend) then measure and so on. So pre programmed for bend. If so, how hard is this?

Robot Build Testing – Togetherness, strength, speed, low centre of gravity, responsiveness, ability to right when (if) gone wrong.

Takeaways – Learn about types of sensors and names, understand terms such as centre of gravity and research videos/information for examples to understand why so important. Analyse for speed – what determines speed (type of motor/quality) or is it power to motor (or both?) Friction on wheels?

December Update

It seems the choice is between sensors that assess distance to walls and maintain within set parameters and/or floor sensor which keeps to the line. With additional knowledge and understanding motors will have to not jolt or lose mms at 45^ turns. Clearance underneath will have to be sufficient for the bump. Any anticipated ‘scrapings’ will be attributable to loss of control and how quickly control can be regained. Must construct testing albeit shorter length of hardboard with rest of parameters the same.