Basic outline of the rules and competition rounds


The bots will all be of essentially identical construction : (example links to suppliers will be provided)

The wooden disk-bot chassis, for which CNC-ready files will be made available.
A pair of MG995 metal-geared servos (rotation mod is required but very simple)
70mm wheels (2 off)
4 AA batteries are allowed - regular AA only, no LiPo or fancy power sources

Other rules :

Sensors, for line follow, edge of the ring detection and opponent detection (sumo) are up to you entirely.
Within any round, the bot must act autonomously, but it may be reprogrammed between rounds.

It has become apparent that there is something “strange” going on with the microbit when trying to use PWM to provide speed control input to continuous rotation servos, the problem has now been reproduced at least three times, so in order to get things moving along ANY controller can be used - within reason ! Arduino, ESP8266 and ESP32 based boards are all fine but there may be a limit on clock speed or something similar if the playing field becomes too uneven through the range of controller boards in use - the whole idea of the competition is to keep things simple and not just throw money as expensive solutions. A Raspberry Pi, for example, costs more then the rest of the bot, so they are excluded.
Lines, and the edge of the sumo ring will be black on a white background and for sensor calibration purposes a test-line is available in the alcove at DoES.

The competition itself will be held over 3 rounds, as below -

Round one = Line follow (the simplest). Very simple to compete, but efficiencies in detecting and making turns will make a difference, as will the sensor spacing and number of sensors. The line will be of fixed width and high contrast, there will be no turns sharper than 90°, there may be curves. This round is a simple time trial, the fastest to the finish wins.

Round 2 = Maze solve (intermediate difficulty). Maze designed to disadvantage simple “side followers” in favour of decision-making. This time, a line must be followed as in round one, fastest to get to the end of the line wins but there may be (multiple) dead ends and T junctions. Once more there will be no sharp corners

Round 3 = Sumo Battle ! A knockout battle with 1V1 competitions in a circular arena of 800mm (tbc) diameter. Start at the edge, detect and then monitor the position of your opponent, finding the right moment to push them completely out of the circle. Two bots enter, one is victorious. The loser must be pushed completely out of the ring.

To ensure fairness and accurate timing of round starts all competitors MUST include a couple of lines of code to watch for a light sensor (which must be fitted on the upper surface) being uncovered. Other than any necessary initialisation of I2c or SPI devices that you may be using, no code must execute until this sensor has been uncovered - this includes no ultrasonic pinging etc.

The plan here is for the timing (probably running on a Pi) and the bot (or bots for the sumo round) to be started at the same time by the simplest means possible - pulling a piece of black card off a light sensor fitted to each fits this bill and will be repeatable within a few mS.

The programming and the choice & positioning of sensors is where the competition will be won or lost - be clever, be sneaky, be whatever you like

What prizes are up for grabs?

Winner in any one round : A single quality street (but not a purple one)

Overall winner : how ever many purple Quality Street there are left in the box.

Plus of course, all the kudos and admiration DoES can muster…

Further details may be added, check back if in doubt.



A quick update :

For testing and development purposes, it is safe to assume that the line follow and maze rounds will use an (approximately) 8mm wide line of black vinyl insulating tape on a white desktop. The tape ordered is declared to be 0.3mm thick.

Where there are “double-back” or parallel sections, the lines will be spaced at a minimum of 100mm (just over half a bot-width) apart. It follows from this that if two consecutive 90° turns in the same direction occur, they will be >100mm apart.

The actual line and maze will be prepared on the day by an independent person, but a calibration line will be available to permit sensor adjustment.