Beyond mileage: Towards more secure techniques to assess the fitness levels of smart cars

Besides its overall optical impression, the assessment of a car's value and/or condition today is widely being based on its (mile)age as a primary indicator. This is a bad and unreliable concept because mileage alone is no representative indicator for a car's condition (which depends on many more fac-tors) and constitutes a focal point for (frequently successful) attacks. In this paper we propose a new approach of determining a more reliable “fitness level” that extends the basic mileage-based estimation. We illustrate advantages a fitness level estimation would yield for different use cases considering security issues like attack resistance. We realize this approach using multimodal, complementary sensor information already present in modern cars. To open discussion with the community we discuss the potential of a first set of 18 proposed properties which include 9 physical, 4 digital and 5 behavior-based ones. Further, this paper proposes a first concept to evaluate the significance of these properties by discussing their explanatory power, freshness, security and available options to verify their plausibility. To a basic extent, this concept could be applied to existing cars, which is illustrated by a practical analysis of a laboratory setup of a 2008 SUV vehicle and a real 2006 limousine car.