Posted in: Cognitive Computing, IBM Research-Ireland, Internet of Things

Hacking an e-bike to help cyclists avoid breathing in polluted air

Everyone knows that cycling is an excellent aerobic sport, but its benefits are less clear when cyclists are riding and exposed to polluted air in dense urban environments. But what if a cognitive technology could help by decreasing the amount of pollution they breathe in?

A project led by Professors David Timoney and Professor Robert Shorten and University College Dublin (UCD), in collaboration with IBM Research – Ireland (Shaun Sweeney and Giovanni Russo) aims to do just this. More specifically, we are instrumenting the e-bikes of the future with technology enabling it to react to real time pollution sensor data and thereby reducing the pedalling effort of a cyclist, i.e. when pollution levels are high, you pedal less.

IBM student Shaun Sweeney was eager to get involved with the project for his Masters thesis because he could see the valuable role that electric bikes can play in transitioning to a sustainable transport energy future.

“We had a great team working on the project from the start, with enthusiastic help from a number of people in both UCD and IBM Research. Within the team, there is a firm belief that smart, context aware electric bikes can provide a number of services that are of value to cyclists in urban environments,” says Shaun.

Hack an e-Bike

 We began by modifying the controller for the motor, added measurement sensors to derive information pertaining to state of battery charge, electrical power input to motor, human power input, and wheel speed.

The bike has also been customized so that it can communicate with a smartphone using Bluetooth, which means that it can exploit external data streams through an internet connection which can then be used to make decisions to control the amount of electrical assistance that the cyclist is provided with. Control inputs are then sent to the bike controller using the same Bluetooth based communication system.

 We then programmed control and optimization algorithms to emulate a natural interaction between the motor and the cyclist recognizing that the natural inclination of the cyclist is to increase effort when the motor engages more speed. In such circumstances we have designed the system to work with the cyclist and not compete with them so the motor switches off and only becomes active when the cyclist is cooperating with the motor.

Finally, we are developing a module for predicting the likely route that the cyclist is going to take using a set of route prediction algorithms that we used in a similar connected car project.[1]

The performance of the system (bike and cyclist) was assessed in indoor tests, where the bicycle was mounted in a commercially available test-stand in order to simulate outdoor route profiles. An important consideration for the control algorithm was to understand how the cyclist’s breathing rate changes in response to changes in the control input. This varies by cyclist, cycling speed and cycling duration and so requires being tested for a range of test subjects.

An e-bike system to mitigate pollution

We see great potential for our context aware e-bike system, particularly with the growth of the e-bike market, which is expected to grow to over 40 million in 2023, according to Navigant Research.

Next steps include trials across of large (and representative) population sample in real urban environments, as well as a theoretical investigation of the cyberphysical control design challenges.

Research Papers

Shaun Sweeney, Rodrigo Ordonez-Hurtado, Francesco Pilla, Giovanni Russo, David Timoney, Robert Shorten, “Cyberphysics, pollution mitigation, and pedelecs”, 2017

[1] S. Sinnott, R. H. Ordonez-Hurtado, G. Russo, and R. Shorten, “On the design of a route parsing engine for connected vehicles with applications”to congestion management systems,”in 2016 IEEE 19th International Conference on Intelligent Transportation Systems (ITSC), pp. 1586 {1591, Nov 2016.

Related Links

Green Transport Spotlight Projects at UCD – Demo Video

 

Save

Comments