Electrifying a wooden bike is as easy as Pi
15 June 2017
Computer scientists at Saarland University modernised the 19th Century ‘dandy horse’, fitting it with an electric motor and Raspberry Pi for pedal-free power.
Back in 1817, German inventor Karl von Drais conceived the “Laufmaschine”, a wooden replacement for the horse and what is known today as a bicycle. This predecessor had no pedals, so the rider propelled it by pushing with their feet.
About 200 years later, computer scientists from the Saarland University made some 21st Century adjustments. Holger Hermanns is professor of computer science at the University of the Saarland and is now well known in the bicycle scene. With his basic research, he wants to help the fast-growing electric bike industry avoid programming errors. In 2011 he introduced a wireless bicycle brake. He demonstrated the reliability of the radio-based brake using mathematical methods, which are also used in control systems of aircraft or chemical plants. The wireless bicycle brake made headlines all over the world and in 2016, the European Research Council awarded him an Advanced Grant worth €2.4 million.
Together with Dries Callebaut, a Belgian bicycle engineer, he is developing a prototype for the "Draisine 200.0". In honour of the inventor, the successor model is made entirely of wood and is braked by a kind of foot pedal on the wooden front wheel. In the hub of the wooden rear wheel sits a 200W electric motor driven by a 750g battery. By cable, the electric motor is connected to a Raspberry Pi which is mounted on the frame of the impeller and is to be controlled by means of an acceleration sensor. Checking the accelerometer and reading wheel-embedded sensors 150 times per second, the Pi activates the hub motor to assist the Draisine, allowing it to reach speeds of 16mph. However, this was not easy.
"In the case of conventional electric bicycles, the engine is switched on when the pedals move, there is no such thing with the Draisine," says Hermanns. Accurately detecting when the driver pushes the impeller is therefore a challenge in fine-tuning the developed control software to the nerve test. The researchers quickly realised that even minor errors have a drastic effect. "Imagine you are bouncing over a curb, the sensor system interprets this as pushing, and accelerates the electric motor to a top speed of 16mph," explains Hermanns.
Many prototypes were tested using a mounted camera on the frame to verify the correct interplay of human and electrical drive power. In order to synchronise the video recordings with the measured sensor data, they also developed a special light-emitting diode clock, which they read out automatically.
The efforts have shown an effect: the third prototype - mini-computer, battery and acceleration sensor have now completely disappeared in the wood frame - can no longer be moved by heavy jogging.
For the researchers this is only the beginning: "We will now verify the correctness of the software, i.e. mathematically prove that the motor does not drive above the permitted maximum speed and that the battery is not overloaded," explains Hermanns. Now, however, the computer science professor wants to take a longer test drive with the current prototype.
For more information, click here.