Introduction

Having watched a number of robot war battles, there are several points to consider when building a robot yourself. To my mind there are four main areas that need to be considered, and within each of these, there are a number of alternatives, each with their own pros and cons. The main categories I think need to be carefully thought through are:

• Shape
  
• Transission
  
• Weaponary
  
• Weight
  
• Control mechanism

Shape

The shapes seem to fall into several classes, some of which give you a tactical advantage, and others that leave you vulnerable to attack. The shell material can also provide advantages or leave you vulnerable. Some of the plastic (or polycarbonate) materials seem to get punctured very easily. Thin aluminum (eg less that 3mm) just doesn't seem to withstand axe attacks very well either. A better cover material would be stainless steel sheet, 2mm thick should do it. I think a fairly rigid triangular based framework underneath the shell is also a must. Even in the shell gets punctured, keeping it away from the delicate components inside will make it survive a bit longer.
No matter what shape you choose, one requirement must be to protect the vulnerable items, such as the radio gear and electronics, in a strong box within the heart of the robot. There should be as much distance as possible from the outside cover to this case so that if an opponents spike penetrates the cover, it won't damage the electronics as well.
I have listed some of my other thoughts on the different shapes below.

Wedges

A wedge seems to give you the advantage of being able to slip underneath opponents and turn them over. It also makes it easier to build a powered scoop to actively raise or turn over opponents. There is a possibility that the driving wheels could protrude both out from the bottom and the top of the robot, which would allow it to continue should it be turned over itself. Failing that, a scoop can act as a self-righting mechanism if it is designed well enough.
No matter what shape is chosen, it is essential that all surfaces of the robot be covered. There are spikes from the arena floor as well as weapons from the opponents to contend with that the robot needs to be protected from. When some of the less well thought out robots have been turned over, you can see all the equipment from underneath where there is no protection at all!

Tanks

A fairly successful body shape has been the tank with caterpillar tracks. They can run either way up (when they are not covered with a mudguards), and if the tracks are of the right type, they can give good grip on the floor. If the tracks aren't designed too well, they can be the Achilles heel instead of a benefit because they have little grip on the floor, or can loose the tracks of the runners.

Others

For want of a better term, I have lumped all the rest together as "others". Provided that the shape gives adequate protection of the vulnerable internal components and will not leave the robot stranded if it gets turned over, then any shape can be used.

Weaponry

There have been countless so called "weapons" used by robots that only pay lip service to the task of causing real damage. There seems little point in having a weapon unless it has weight as well as speed with which to inflict damage. Hypnodisk was one of the few that had real damage capability. It had such a colossal flywheel that span at such a high speed, that the damage it caused was considerable. Having swinging hammers on a rotating spindle seems to offer little benefit when it can be stopped after the first impact with the opponent. You need a weapon with high momentum and a short "bite", and one that doesn't come to rest after the first blow. You don't want a large bite on a rotating weapon because it's a bit like moving a lathe's cutting tool too quickly into the workpiece. You need to do it a bit at a time to get the best result. E.g Hypnodisk works well because of this.
If you want an axe type weapon, it needs to have weight at the extremity, a sturdy shaft, and a high-speed action. The best axes are hydraulically operated like Razor, and the house robot Shunt, which both seem capable of awesome damage. Pneumatic rams can be good, but you need reasonably sized units, and maybe more than one ram per weapon to give it real power. There have been some weapons that are activated by starter motors. These seem to have the speed, but the ones I have seen appear to come to rest too quickly on impact. Maybe with a bit more weight they can compete with hydraulically operated units

Weight

There are several different weight categories in Robot Wars. The ones you see on Television are the heavy weights. The program producers also choose robots that look appealing to the eye, are more impressive, and will therefore attract more viewers. There are also lightweights, which look more like the traditional radio controlled car racers. I think it will be more interesting and challenging to build a heavyweight robot, and in fact it may be difficult for it not to be a heavyweight by the time you have batteries, gas cylinders, electric motors etc.,under the bonnet. Weight also allows you to be more aggressive and not be pushed about too much.

Transmission

There seems to be only two alternative for transmission types: electric and internal combustion engine (IC). If you choose IC, you need to have a gearbox, clutch, etc, which is more complicated, heavy, and required more room. Also you need to consider the fuel situation if you get turned over. Electric motors are cleaner, don't require a gearbox or a clutch, but do require electric batteries, which are heavy.
Consideration also needs to be given to the final drive method. E.g. do you use caterpillar tracks and if so are they metal or rubber? Do you use wheels, and if so what type? How many wheels do you include and how many need to be driving wheels? On a recent Robot wars episode, a four wheel robot easily pushed a caterpillar tracked robot sideways across the arena floor and into the pit. You might have thought that the surface area of the caterpillar tracks would have given enough grip to prevent this, but no, the rubber tyres were better. They looked like go-kart "slicks".
There is also the issue of steering. Do you go for turning the wheels, or simply driving one side faster that the other? Turning wheels adds more complication and a potential for being damaged. Speed control of each side scrubs the tyres, but is a simpler solution.
Braking needs to be considered and can be included in two ways. The conventional method with disks or shoes, but this can be complicated and expensive. An alternative, if you use electric motors, is to use electric braking, sometimes called power regeneration. It even comes as a by-product with most types of motor drive systems. You effectively short out the motor terminals when it's not being driven!
I think whatever transmission method you use, it must provide enough grip and power to shove an opponent across the arena floor. Speed gives you added momentum for attacks, and traction gives you the advantage of being able to shove other robots about, as well and prevent you being shoved about yourself.

Control mechanism

The issues that immediately come to my mind are how to interface the radio receiver to the robot controls such as the motor control and weapon actuators.

Some people have used a very crude but simple method such as that show in the adjacent picture. Here the speed control joystick within the robot has two servos connected to it, which then physically move the joystick as required. To me, this makes the system vulnerable and prone to breakdown. It is also possible for the servos to become dislodged and render the joystick inoperable even though the servos themselves may still be functioning okay.

My suggested alternative is to use a small microprocessor board that connects directly to the receiver. These boards can be bought commercially and have the added advantage of reducing the number of superfluous components. Having given it a bit of thought, the processor board can take the place of the servos in the above picture because it can perform the speed control function of the motors itself. It can therefore remove the need to have the joystick, as well as reducing the hardware needed in the associated speed control circuit.
If you didn't know already, there are "cheap and dirty" and "sophisticated" ways of providing motor speed control. The cheap and dirty ones are basically on/off switches! The more sophisticated ones (which are the ones I would favour) are "chopper" types that create pulses of varying width and/or frequency, and allow a very fine control of the motors. They also do not "waste" power if you used a "loss resistors" type speed control method which simply throws the power away in heat.
There have been numerous occasions when robots seem to "die" just as the bout starts. It is not always obvious why they die, but I can guess that it must be partly due to the interference of the metal surrounds, the other robots, and internal interference from their own motors. However, if the radio gear is shielded sufficiently inside the robot, and suppression is sensibly applied to your own motors, I think this should not be too much of an issue. I also wonder if it is because of the mechanical nature of the servos and what they are trying to control. The servos are small and only meant to operate light actuators like those in model aircrafts. When they are used in a robots, the environment is a bit more hostile.

Summary

After carefully considering the various areas listed above, I think the first draft Robot design should be as follows:

• Four rubber wheels to give maximum traction on the arena floor.
  
• Electric driven transmission.
  
• One motor to drive both wheels on each side
  
• Chain and gear reduction from motors to wheels
  
• Steering via speed control of wheels on each side.
  
• Triangulated metal framework internally, covered with 2mm stainless steel shell on all surfaces.
  
• Wedge shaped front end with twin ram pneumatically operated flipper at front that doubles up as a self-righting mechanism.
  
• Pneumatically operated swinging axe with as much weight as possible at the extremity, and which has a sharp point.
  
• Radio receiver integrated with Microprocessor based controller to reduce the number of mechanical servos needed.
  
• Speed control handled by microprocessor board.
  
• Axe operation handled by microprocessor board
  
• Motors and all electrical items to be adequately suppressed.
  
• Vulnerable internal components to be housed in the centre of robot in a metal box.

Components such as the motors can either be bought new (a bit expensive especially if the robot gets obliterated in the first few seconds) or from surplus stores or scrap yards. Could use motors from wheel chairs, lawnmowers, forklift trucks, invalid carriages, washing machines, etc. Car starter motors are NOT very suitable because they are "series field" motors and are very difficult to speed control. Other components can be obtained from various sources or scrounged from friends.