jawfish

Toys are Better Robots
john posted in technologies on June 22nd, 2005

The Penguin school vs the Corporate school. Why hobbyists and toys sometimes make for better engineering.

Here are two robotics examples-

Penguin school- ChRoMicro an off-the-shelf helicopter home project with an off-the-shelf cpu board:

This is an alpha level attempt to build an autonomous helicopter. It uses a tiny RC electric indoor helicopter, a laptop, an Xbox game controller, and a gumstix tiny cpu board with bluetooth networking. This is all equipment available to any hobbyist.

Good points:
It has a very sophisticated on-board cpu, a well-developed communications protocol, Bluetooth, an ergonomic controller.
The vehicle is an electric helicopter, with short range.
Using on off-board computer to do the command decision making is a fine way to optimize resources. With a dead-man circuit on the helicopter, this design could make a commercially-reliable autonomous robot platform.
This adds up to a mature platform with many tools available, at extremely low cost for a prototype.
The project is easily scaled to a larger helicopter with greater range, or any other RC vehicle like a plane, a boat or a car..

Critique:
Bluetooth has very limited range, ChRoMicro plans to move to WiFi, which is also limited, but maybe good enough.

Corporate school- a system of robotics:

Cyberotics provides multiple platforms and controllers, with add-on hardware.

radio-modem – wireless connection

This is a sophisticated wheeled platform with plug-in capability for sensors. The radio kit enables simple remote computation and the on-board cpu, PC104, is an industry standard item. The same company sells WeBots software – a design and simulation environment that looks very interesting.

Good points:
Standard cpu board, good-looking exterior, sensors look high-quality. Software has great promise.
The software is scalable, but the vehicle hardware is not.

Critique:
The radio link is RS232, which is less sophisticated than Bluetooth or WiFi, though it is well supported. It is much too slow to support an autnomous robot controlled by an off-board computer.
Cost is exorbitant-
WeBots software – 300 or 3600 chf $240 – $2800 USD
Koala platform – 7260 chf $5700 USD
Radio modem – 2750 chf $2160 USD

Conclusion:
Both examples are interesting, usable platforms that look to be capable of real robotics research. The commercial software looks great. There is more computing power in the ChRoMicro project due to the off-board laptop. It is much more scalable and flexible. The ChRoMicro project is also capable of autonomous control thanks to the broadband wireless link. The cost of the commercial kit is on the order of $10,300 -$13,600 USD. The ChRoMicro parts, less laptop, cost less than $1,000 USD.

Traditional corporate engineering can produce excellent results, but the man-hours and overhead create a 10x increase in cost. If the market warranted, the toy company that produces the RC vehicles could add the wireless digital link very cheaply. They already sell camera kits for RC aircraft. Commercial software with the inconvenience of licensing, with costs over a few hundred dollars will never be widely used, and is an easy target for free Open Source duplicates.