The Mbed Robot. It is  RC Tank, with the electronics, turret and upper hull removed, motors left in. Four IR proximity sensors have been mounted on 2 servos, and a lead acid battery provides the power. The goal is to make the robot drive around without crashing. Simple... ish. Later goals include doing intresting things,

• finding its way from point A to point B, 
• operating on a time lag,
• dynamic mapping of an enviroment, whilst talking to a command station (laptop, command station sounds better).

This bit has absoulutly nothing to do with the mobile robot project that I have to do for my degree. Thats done on PIC's, subtle but important diffrence.

• Development of a motor driver libary is ongoing. The motor driver hardware has been finaly debugged. The code has been tested, using a set of logic switches that I know work, and the actual motors.

• Some playing around with the range finders is required to calibrate them. A libary to play with these has been written. Testing proves some relitivly accurate returns. Not brilliant, but good enough.

• Servos, Im using the existing libary. Intresting bits are going on here, I just ran the hello world, and now need to calibrate them so they do what they are supposed to.

• Batery monitering, kinda low on the priority list, its a lead acid battery, its going to take a while to make that flat. I can multimeter it till I work this one out. An idea exists to use switching to minimise power consumption of a potential divider, and then mesure battery voltage.

• Paint scheme, as this should look cool. or at least wierd, which it does quite well.

• Anything else I can make it do.

As of 25/11/10

• Working motors, not running at the same speed, but they both work. Speed issue is due to cheep chinese motors, you get what you pay for. Encoders are needed, so some thinking will go on here.

• Working sensors, results could be more accurate, but are good enough. I will have to calibrate them properly later.

• Working servos. Not really used yet.

• Working object avoidance. Well, if its not a chair leg... and some minor glitches like been attached to a power cable, not seeing anything behind it, et al. Youtube video will be done when the robots charged.

• Now to charge the battery. I have a workaround for charging the battery, crocodile clips attached to the charger leads. It works, duct tape prevents anything shorting on the battery conectors.

As of 28/11/10

• A single small strand from the multicore cable used on the motordriver shorted with the capacitor. Fifteen minutes wasted trying to find that. Need a PCB for the motordriver.

• Object avoidance algorthim was written in 5-10 minute based on that for small three wheeled robots. This is a big tank, hence some more minor glitches. Like a masive turning circle. Oh and small fluffy teddy bears just get squished, sofas are fine though.

• I guess infrared returns from fluffy objects are scattered, and therfore not providing a strong enough return to be picked up. Also anything below the hight of the sensors gets ignored though it can drive straight over most of that.

As of 29/11/10

• Implemented a logo based set of functions. Provides some intresting options for the next set of object avoidance algorithms.

As of 1/12/10

1. low speeds are impossible due to the mass of the robot, particuly when turning.

1. letting the robot drive at full speed for 0.4 seconds is stupid.

1. it is possible to connect the motor the wrong way round.

• changes involve slowing the robot down and decreasing the time spent between checks on surroundings.

• The results are pretty good. It avoids objects it detects and runs over the rest. Unfortunatly it can easily get stuck due to its size.