The Bus Overtake

You are an autonomous bus with sensing capacities and steering algorithm. You are just a little bit smarter than an older generation bus, which enables you to overtake. Carry out an overtaking manoeuvre, but don’t stop the traffic flow!

Preparation

For this game, you need a wooden or metal frame of the size of approximately 2.4 x 1.5 meters which is the body of the bus, equipped with distance sensors in each corner.

You also need an autonomous bus. Sorry, we know that this is not always the easiest to come by, but it seems those autonomous buses drive around in ever-increasing numbers.

Toy Bus Specs

Dimensions: 2500mm / 1400mm

Material: box section aluminium 55mm / 55mm with 2mm walls.

Construction: The corners were all mitre cut Each section was joined and reinforced with wooden batons bolted together with steel corner brackets (see images).

Sensors The four project boxes with the sensors and led-display must be fixed to the four corners of the frame (for example with self-adhesive velcro) so that sensors point outwards extending the four corners.

The sensor boxes The role of the sensor boxes is to measure the distance between objects in the environment and the corner of the frame and to provide visual feedback about the distance. In the game release setup following parts were used: ToF sensor Arduino uno (or other Arduino with SDA and SLC pins needed for the I2C communication) Logic level converter (make sure they are bidirectional) Power bank (between 1 and 2 Amper, 3.5 - 5 Volts) Neopixel LED strips (>=16 LEDS)

The sensors are directed outwards in an angle of 135º from either side of the frame and measure the distance from the frame in a range of 0-20(?) meters. A led light row reflects distances between 0-5 meters by turning from green to red gradually when the distance to a nearby object decreases.

Calibration Make two marks on the led scale mark ‘close’ and ‘too close’. We chose to calibrate our system at +- 2 meters for ‘close’ and +- 50 cm for ‘too close’.

Code on Github

There are two code bases on github First is the tiny web app that allows players to play as AI trainers and sensors https://github.com/free-ice-cream/catch-the-bus

The second repo is the code for the Lasers sensors and the neopixel display. https://github.com/free-ice-cream/catch-the-bus-sensor

Rules

Five players is an ideal number for ‘bus’ players (with a minimum of three players). ‘Bus’ players are carrying the body of the bus and are walking with it ‘Fred Flintstone’ style.

‘Bus’ players consist of one ‘bus algorithm’ player and ‘bus sensor’ with the remaining 2-4 players. ‘Bus sensor’ players occupy each corner of the body equipped with a sensor box. If you have less than four ‘bus sensor’ players, make sure to cover the front end of the bus first. The ‘bus sensor’ players main task is to pay attention to the distance readings on their sensor box and report to the ‘bus algorithm’ whether there are any ‘close’ or ‘too close’ objects. The ‘bus algorithm’ player is blindfolded (listening only to the ‘bus sensors’) and steers the bus being positioned in the back of the body.

The ‘toy bus’ positions itself behind the autonomous public transport bus.

A non-playing character positions themselves in front of the autonomous bus. The game starts when the non-playing character starts to walk slowly, so determining the speed of the autonomous bus (which can not overtake and thus adapts to the walking speed of the non-playing character).

A ‘toy bus’ now needs to carry out the overtaking manoeuvre. When the ‘toy bus’ comes too close to the autonomous bus it will stop, based on the actual settings of the specific bus. The ‘toy bus’ has to overtake with bringing the autonomous bus to halt no more than three times.

Players win if they succeed in overtaking the autonomous bus without bringing them to stop no more than three times.