How the Agoria Solar Team ensures its place in the sun with AI

The SASOL Solar Challenge was originally scheduled for September 2020. Due to the Covid-19 pandemic, however, the race was postponed to February 2021. “In normal circumstances, all 25 members of the team work together in our central workshop to strategize, design and test,” says Romanie. “Like many people, we were forced to work remotely and adapt our plans accordingly. But the goal remains the same, of course: winning the race.”

Planning for a victory lap

In that department, the 2020-21 Solar Team has big shoes to fill: their predecessors won first place in the Carrera Solar Atacama in Chile (2018) and the Bridgestone World Solar Challenge in Australia (2019). As strategists, Romanie and Kenneth need to come up with a solid plan that will guide the team to victory once more. For this, they build on the experience of past teams and accurately predict the impact of environmental factors to allow in-race adjustments. 

In any case, a simple copy-paste of the 2019 strategy won’t cut it in South Africa. “It’s a completely different race,” explains Kenneth. “In Australia, the goal was to travel a fixed distance in the shortest time possible. In South Africa, it’s the other way around: the time is fixed, and the goal is to complete as many laps on a set course as possible. That premise requires a completely different approach.”

The brains behind it all

During the race, Romanie and Kenneth will be directly behind the BluePoint solar car in another car, which they affectionately call ‘the Brains’. From here, they provide the pilot with instructions – speed up, slow down, overtake, don’t overtake, etc. These instructions are based on the broader strategy, but are also determined on the fly by data and predictions provided in real time.

“Our solar car, BluePoint, is equipped with two very basic sensor points: a speedometer and an inclinometer,” says Romanie. “That’s done deliberately to keep energy consumption on the race car as low as possible. Most data capture is actually done directly from ‘the Brains’. Here, we monitor a plethora of factors, like energy consumption, power, weather conditions, etc.”

Wind, power, and machine learning

As the Innovation Partner of the Agoria Solar Team, delaware played an important role in helping the team make sense of the data and glean actionable insights that could be used to optimize performance. “Together with delaware’s AI team, we developed different models for power consumption and how this is impacted by environmental factors and speed,” Kenneth explains. “Since we will be driving in circles this time, the interaction between acceleration and consumption is something that needs to be accurately measured in order to be successful.”

During the race, wind power plays a key role as well. “BluePoint is designed in such a way that we can take advantage of the wind to achieve a ‘sail’ effect,” says Romanie. “For this, we need real-time data and accurate prediction models that allow us to give the pilot the right instructions. Here as well, delaware’s experience and expertise with AI and machine-learning models proved invaluable.”

Bringing it all together

Romanie and Kenneth’s work isn’t limited to interpreting and acting on data from sensors, however. As strategists, they also need to take other, less obvious factors into account, like training, fatigue, how long it takes to change drivers, etc. Romanie: “Strategy is where it all comes together. We need to constantly make simulations of how certain factors will impact our performance and chances of winning. It’s a challenging task, so it helps to have the delaware AI team on board who thinks along with us and brings new and innovative ideas to the table that expand our idea of what’s possible.” 

Follow the Agoria Solar Team at www.solarteam.be or learn more about our AI approach at delaware.ai.