Swiss Re Coloride: Understanding driving behaviour to reduce risk
From protection to prevention
Insurance plays an important part in people’s lives – helping them to protect their loved ones and properties, and enabling them to claim damages and receive financial compensation, when something goes wrong. More and more insurers are shifting their focus from protecting people and properties, to risk reduction and prevention, with new technologies helping to make this shift possible. Devices like smartphones and smart watches accompany people everywhere, giving them access to a range of services and connecting them directly to different service providers, but also by helping to collect data and providing insights into people’s behaviour.
Driving behaviour is a good example: it is now possible to collect concrete driving data through people’s smartphones. This data can be used to study risky behaviours and enable companies to develop plug and play solutions for their customers, alerting them to risky manoeuvres or providing other solutions to prevent accidents. This is one of the focus areas of Movingdots, Swiss Re’s technology hub in the automotive and mobility space, in developing applications that combine advanced technologies with an intuitive customer experience, essential for success in this area.
Swiss Re’s telematics app, Coloride, has been developed to bring risk reduction and prevention to the field of motor insurance. Coloride can be easily installed on a smartphone and is able to analyse driving behaviours and raise awareness about dangerous driving conducts. The app helps drivers prevent risks, through Coloride’s detection module, developed to spot dangerous driving habits like harsh manoeuvres and phone distractions.
How Coloride analyses driving behaviour
A core component of Coloride is the detection of driving habits, which make up a person’s driving behaviour, like certain manoeuvres and phone distractions that increase the risk of accidents. Figure 1 shows the steps, from the app detecting certain driving manoeuvres to providing insights to both the driver and the insurance company.
Figure 1. Coloride: from detecting driving risky behaviours to informing the driver and insurance
Coloride can identify several driving manoeuvres, such as acceleration, braking, cornering, harsh steering, driving around roundabouts and at junctions, U-turns, all based on raw accelerometer-data and GPS signals collected directly from embedded sensors in drivers’ smartphones.
These data are then processed, using standard and novel signal-processing methods to remove noise and transform it into useful information that describes the car’s motions. Swiss Re’s Advanced Analytics team has developed proprietary algorithms to align phone accelerometer signals to the car reference frame, so that they can be used in determining the vehicle motion dynamics. This is a crucial step in the process, because drivers are not asked to put their mobile phone in any type of holder to keep it stationary when driving, so the algorithm must distinguish all the different orientations that the phone could assume during the trip and apply the proper corrections to align its signals with the vehicle’s motion frame. In addition, GPS signals are enriched with contextual map content from Swiss Re’s partner HERE. This helps to detect and further characterise manoeuvres such as harsh braking in front of a pedestrian crossing or near a school zone.
Three steps to detect and process risk-increasing driving manoeuvres
Coloride’s detection module follows three main steps to detect and process data in parallel across different manoeuvre types (see Figure 2) the first step potential manoeuvres are extracted from the signals of GPS-derived features (acceleration, braking, cornering), accelerometer-derived features (harsh steering) and contextually-based features (roundabout, junction, U-turn). Once a potential event is extracted, the GPS-, contextual- and accelerometer-derived features are computed to characterise it, as part of the second step. These features are used by machine learning models in the ruleset evaluation step to predict if a potential manoeuvre can be released to the third and final step. This evaluation step is necessary to remove potential non-harsh events, as the previous step can generate false positives, and its aim is to generate candidates that comprehensively cover all possibilities of actual harsh events i.e. prioritising recall performance instead of precision.
Machine learning models are used to filter out “bad” candidates in the ruleset evaluation step. These models are trained against data, collected during both on-track and on-the-road car tests. The on-track data is used to learn signal patterns characteristic of each manoeuvre type, while the on-the-road data helps to finetune the models on realistic car trips. For this purpose vehicles are equipped with devices that can collect high-quality accelerometer- and GPS data. These data are used as a reference to train algorithms in dealing with the lower quality signals that are typically collected from smartphones.
In the final stage, called post-processing, manoeuvres from the previous steps are put together, using both disambiguation and contextual map-based validation logics to produce unified and coherent output.
Figure 2: Coloride’s detection module: 3 steps to detect and process risk-increasing driving manoeuvres
Detecting phone distractions
The detection of phone distractions, which is also part of Coloride’s detection module, is based on events available in the smartphones’ operating system and on events and features that are computed based on accelerometer signals. The locking and unlocking of a phone are used as possible start and end triggers of phone distractions. They are considered in combination with phone handling events computed from the accelerometer signals. These two sources of information are used to detect phone usage while driving, e.g. reading notifications, texting and calling.
Insights into risk increasing driving behaviour for drivers and insurers
Once a trip is completed, the output of the detection module is displayed on Coloride’s map. Together with speeding and contextual information, these form the building blocks of Swiss Re’s Advanced Scoring services, which help evaluate the behaviour of policyholders behind the steering wheel from an insurance perspective. These insights can help drivers to better understand their driving behaviour and how they can reduce their risk (behaviour steering).
Figure 3: Insurer’s dashboard insights Figure 4: Harsh
It provides insurers with valuable insights into their customers’ driving behaviour, to enable risk-based pricing. Risk-based pricing and behaviour steering are two of the five telematics value-creation levers, together with risk selection (try-before-you-buy propositions), value-added services (cross-selling other insurance products, such as travel insurance once you drive across the border) and loss control (first notice of loss and other loss mitigation measures).
Swiss Re’s Automotive & Mobility Solutions and Movingdots help insurers to develop and implement their telematics programs, focusing on a single lever or a mix of different elements, based on their portfolio, innovation appetite and reference market.
About Swiss Re
The Swiss Re Group is one of the world’s leading providers of reinsurance, insurance and other forms of insurance-based risk transfer, working to make the world more resilient. It anticipates and manages risk – from natural catastrophes to climate change, from ageing populations to cybercrime. The aim of the Swiss Re Group is to enable society to thrive and progress, creating new opportunities and solutions for its clients. Headquartered in Zurich, Switzerland, where it was founded in 1863, the Swiss Re Group operates through a network of around 80 offices globally.
In 2017, Swiss Re created its P&C Solutions unit, which aims to help clients improve their profitability, increase efficiency and unlock growth, by enabling it to make better underwriting and pricing decisions, using its range of technology and data analytics tools across various P&C lines of business.
The Automotive and Mobility Solutions (AMS) team at Swiss Re leverages its expertise in data analytics and its strong network of strategic partners to design and create innovative products addressing the key challenges of primary insurers. The AMS product suite includes tailored insurance cover for specific electric vehicle features, including end-to-end, modular telematics that enable the assessment of driver behaviour. It provides a complete claims settlement platform and vehicle underwriting platform, which allow accurate assessment of vehicle safety features (ADAS Risk Score).