Campus and Community

CycleAI: the app that will help cyclists to feel safer

The app was developed by the Técnico alumnus Luís Rita, who was recently distinguished as one of the “Top Talents Under 25".

Imagine yourself leaving your house for a bike ride with friends on a sunny Spring day. You are fulfilling all the rules for cycling, and you end up being run over while cycling on a bike path. It happens more often than we might think. It happenned with Miguel Sampaio Peliteiro, a doctor who was violently ran over while cycling on a bike path in Póvoa de Varzim, in May 2020. Miguel Sampaio Peliteiro spent three weeks in a coma following the accident, he was hospitalized twice and had a hard rehabilitation process that extends until today, with some improvements along the way, but also sequels. The young doctor started walking and cycling again, and he will do everything to ensure that this situation is not repeated.

Miguel Sampaio Peliteiro heard about Luís Rita’s master’s thesis related to mapping of cycling risk, and he thought he would be the perfect ally. “He sent me a message saying that I couldn’t leave my master’s thesis ‘Become a Better Cyclist with Deep Learning’ on the shelf – he told me that the project was great, he wanted us to take it further, and turn it into a real-world project”, recalls Luís Rita.

“The day after we spoke for the first time, we were already meeting with national leading entities in the field of mobility and urban planning”, says Luís Rita. The commitment and dedication of the two entrepreneurs resulted in CycleAI app, developed by Codrin Bostan.

Using artificial intelligence algorithms, CycleAI intends to give more power of decision and safety to cyclists, making urban planning more efficient and analytical. “By studying the risk perception from the individual’s point of view, we aim to enhance safety in cities, in which the consequences of accidents involving cyclists and pedestrians tend to zero”, highlights Luís Rita.

Just a few months after the beginning of this adventure, the two entrepreneurs ranked 1st in the inaugural ECF hackathon and more recently in the BET / AWS University Challenge. “It was amazing. It was our first competition. We were fully convinced that our project could make a difference in people’s lives and in the way cities develop”, says the Técnico alumnus. “Winning the hackathon encouraged us to follow our work”, he also stresses.

The creators of CycleAI made available on the website a map where users can mark areas that they perceive as risk zones, national and international. “It is a crowdsourcing that articulates a world map of hotspots with the most dangerous places for cyclists”, explains Luís Rita. Cyclists can place a pin or write an address, adding a description that justifies their classification or add an image of the place. The map will always be available to the public, thus allowing the app itself to keep up with the rapidly changing cities.

At the moment, Luís Rita and Miguel Sampaio Peliteiro are already developing a second crowdsourcing, which is different in its purpose, since the data collected will allow training the machine learning model so that one can interpret the sense of risk perception of an image in a matter of milliseconds. “We will ask users to choose between two images caught on Google Street View the one that gives them more security”, explains Luís Rita.

Despite the investment made in prevention, cycling mobility must be safer

The truth is that although there are an increased number of incentives for cycling, in many cities, particularly in Portuguese, is still not as safe as we would like. According to a recent study by the European Cyclists’ Federation (ECF), Portugal is one of the European countries with worst conditions for cycling (ranked 27th among the 28 countries evaluated as bike-friendly). “High accident rates despite the investment made in prevention” is one of the causes referred in the study.

The data collected by CycleAI can be used in many ways, but the main goal is to give cyclists more tools to stay safe. “Centralized data processing will increase not only cyclist safety but also those who are part of the urban ecosystem, whether pedestrians, motorcyclists, car drivers”, stresses Luís Rita.

The Técnico alumnus also recalled that active mobility has become an important focus for European municipalities and governments in recent years: “Adaptation of cities for this purpose is increasingly urgent, namely building data hubs that allow to ensure the safety and well-being of society in general”.

The project is achieving “superb visibility and feedback, well above our expectations”, says Luís Rita. The Técnico alumnus is aware that people are often reluctant until projects become tangible. Even so, the two entrepreneurs are “confident that, as soon as we establish public and institutional agreements, we will be able to transmit this feeling of accomplishment”.

The first data collected by CycleAI should be known in early 2021, when the application prototype should also be ready. Luís Rita and Miguel Sampaio Peliteiro were invited to participate in Velo-City 2021, the main world conference on cycling mobility. They accepted the challenge, and they will cycle from Barcelona to Lisbon, while testing their final product. “It will be really amazing”, says the Técnico alumnus.

“Above all this is the materialization of a work that was merely academic”, says Luís Rita, “Turning my ideas into real-world applications, is something that I really want to do”, he adds. “CycleAI aims to save and improve lives, and we intend to continue to work with that in mind”.