Collecting machines’ sensor and production data, predictive maintenance anticipates incurring failure. It plans maintenance accordingly such that downtime and repair time can be minimized or even avoided. What sounds like a great innovation for car owners, however, raises some questions and fears for the future of independent car repair shops: Will this develop into a catastrophe or an opportunity?
Relying on latest technological progress, predictive maintenance can be established in the field of telematics - an interdisciplinary area which combines vehicular telecommunication and computer science. It goes without saying that the automotive industry is a vital use case for predictive maintenance: Cars gather data from the individual parts using the connectivity of the Internet of Things.
You might hesitate here: Hasn’t this practice been on the market for decades, for instance cars warn about the next engine check in regular time periods? However, this refers to preventive maintenance: The provisional regular replacement of pieces without considering the actual state of the parts. Subsequently, parts in excellent condition happen to be restored and failure candidates are not detected soon enough.
In an ideal world, predictive maintenance detects a tendency for malfunction before it occurs such that a break-down never happens. How can one ensure this precise prediction? Three points are key here:
- Data collection and transmission. Big data is the enabling factor: High amounts of data need to be tracked and exchanged to be able to contemplate all dependent variables for a correct prediction. This could surface unforeseen correlations: Values from odometers and repair history are considered, but also temperature and air humidity measures are part of the feature vector that is later fed into the predicting algorithm. To handle the high amounts of data, big-data techniques such as Edge Computing are taken advantage of. Currently, most of the vehicle data is directly sent to and processed by the car manufacturers.
- Storage and analysis of data. The more data, the better: A high prediction accuracy can only be achieved when all influencing factors are monitored over time. Thus databases are demanded to provide a high capacity. The data is analysed with state of the art AI techniques where supervised and unsupervised machine learning methods have proven to be successful.
- Calculation of occurrence likelihoods. Having collected patterns of the metrics in certain states, a mathematical vector model can then compare similar instances. If we refer back to our previous fuel pump case, the likelihood of a soon failure could rise when the attributes of the fuel pump significantly change.
Predictive maintenance ensures an optimal cycle of spare parts replacements. What does this imply for car repair shops? At first glance, it seems that independent repair shops will lose this game, since the car manufacturers own the vehicle data and the car will directly be handed to a manufacturer’s car repair shop in case of emergency. Another first thought is that the vehicle will need less regular maintenance checking, thus customers will need service less frequently.
However, we take a second glance: Following the data protection trends within the automotive market in the EU, stricter laws are laid upon OEMs. Looking into the future, we believe that the EU will enable an Open-Telematics-Platform (OTP) that will provide fair and open reading access to sensors, control units and other data. This development is already in its infancy: Open telematic platforms are part of the Europan intelligent transportation systems (ITS) action plan and ITS-guideline.
Aside from political entities, companies are already working on open data sharing solutions in parallel. Caruso or Carmunication for instance are providing multi-brand in-vehicle data of different vehicle manufacturers in one location.
Consequently, independent car repair shops will receive more and more access to the data and can connect directly to the customers’ cars.
We also keep in mind that the possible lifespan of the vehicle components is not affected by predictive maintenance, thus it will not significantly affect the overall market demand for spare parts.
Even though car repair shops will have to adapt to predictive maintenance, change brings opportunity: Repair shops will be able to plan ahead through predictive maintenance and focus more on quality and service which is growing in importance for customers and therefore be able to keep their clients. The quick delivery of ordered components will not be a limiting factor anymore since the demand is communicated in advance.
How will workshops need to steel themselves for the incoming changes? One requirement for independent car repair shops is the investment in high quality CRM systems and efficient vehicle communication to realise predictive maintenance. Since independent repair shops work on a high variety of cars, they can collect more general data than the brand dependent workshops and have more experience with older car models which could be used as a niche. According to BearingPoint, younger drivers prefer independent repair shops over OEM workshops. Young people also tend to be most open towards new digital solutions. This combination is a chance to win customers by quickly adapting to digital change in the form of predictive maintenance.
At auteon, we embrace digital development and already vision about incorporating predictive maintenance car sensor data into our software and hence suggesting the perfect spare parts. We want to contribute to the future aftermarket by providing a transparent and fair way of sharing vehicle information that supports independent car repair shops. We see an open telematics platform as a must to guarantee fair conditions for all market players and are ready to step up to reach this goal.
In a fast moving world, we thrive towards the latest technological standard and welcome new innovations. We stay on track with the industry progress within the field of telematics and are awaiting the day when the first car AI wants to use auteon.