The human brain is keyed to narrative. Inputs, outputs, historic events, future possibilities, correlation, and causation: we are wired to understand the world around us as a constantly unfolding story.
A Formula 1 race is also a story in progress – and one in which we can proactively shape the outcome in our favour if we take intelligently informed action at the right time. Choosing that action, and the right moment, is the key.
That’s why McLaren Applied Technologies developed ATLAS (Advanced Telemetry Linked Acquisition System), an advanced data-visualisation tool that can provide powerful, race-winning insights.
Every Formula 1 car carries hundreds of sensors, monitoring every performance parameter from suspension movement to engine temperature, and transmitting mission-critical data in real-time to every team. Over the course of a grand prix – or indeed, any test, practice or qualifying session – that amounts to millions of individual data points.
Combine that with other relevant data, such as the location and speed of the car on track compared with every other car, and you have a vast wealth of information.
By presenting performance-critical data in a form that’s sympathetic to the human brain’s preferred means of interpreting it, ATLAS enables engineers and strategists to maximise their abilities. Expressing information contextually, and as patterns rather than plain numbers, enables experts to understand the correlation and relationships between the many pieces of data as they arrive in real time. They can immerse themselves in the data without being overwhelmed by it, assess the risks versus the possible outcomes of any strategic call, or rapidly identify any performance parameters that diverge from historic trends. If a component is showing signs of unexpected failure, we can instruct the driver to stop the car and prevent greater damage.
ATLAS can also overlay current data with historical data, helping to build an understanding of the car’s development trajectory – whether there are improvements, where they are, and how they compare with the results of simulations. Are setup changes or new components yielding the anticipated results? If not, why? Whether the outcome of any change exceeds or fails to match simulation results, it’s vitally important to understand the reasons. Motorsport is, in effect, a simultaneous science experiment performed under pressure and with limited resources; failures can be challenging or even impossible to reproduce in controlled conditions, which makes live data acquisition crucial in understanding not just the symptoms, but the causes of those symptoms.
McLaren Applied Technologies first developed ATLAS at the turn of the 1990s for customers in Group C sportscar racing, a fuel-restricted category in which pitstop strategy was crucial. We have evolved it beyond recognition since then: a system that initially required an entire truck to itself can now operate on a Windows PC while processing a greater volume of data than ever before, and is used by clients in a broad range of top-level motorsport categories.
Data can be stored in any SQL server, so compatibility is ensured, and ATLAS is customisable to suit a variety of applications and user preferences, whether the end user is an individual or a group. ATLAS 10, the latest build, offers an even richer set of predictive tools, and enables users to monitor the accuracy of predictions in real time, opening the door to ever-greater simulation fidelity.
Beyond motorsport, ATLAS is an ideal tool to understand and manage any complex systems that have many degrees of freedom.
As the automotive industry looks ahead to a world in which connected and autonomous vehicles become the norm, real-time performance monitoring and sophisticated, highly accurate predictive algorithms will provide the bedrock of customer safety and satisfaction. ATLAS will help navigate towards this future.
In the public transport field, for instance, if a train begins to fail while in service, the operator needs to understand the causes and implications quickly. Is it an isolated fault? And whether or not that is the case, what effect will this event have on the running schedule across the entire network?