Research On Data Update Method In Vehicle Electronic Environment

With the rapid development of automobile electronic technology, a large numberof sensors and electronic control units are widely used in modern vehicles for security,comfort and intelligent services support. Vehicle electronic control system is a typicaldistributed real time system. It needs to schedule the real time transactions fromsensors and electronic control units and manage a mount of real time data withdistributed real time database. The real time data update is limited by validity period,so the real time transaction must update its data before deadline, or the data are notavailable. When traveling on the road, the vehicle can communicate with each otherfor data sharing by vehicle ad-hoc networks (VANET). Thus the vehicle controlsystem can well obtain the traffic environment information. The accident can beavoided after an early decision when emergency situation happens. However, thewireless network is unstable, and the VANET is different from mobile ad-hocnetworks (MANET). The data update of mobile real time transactions should be forfurther research in vehicle wireless communication environment.In vehicle control system management, the real time data can be divided intobase data and derived data. The base data and derived data in the same transactionhave dependence relationship, which can form a directed acyclic data dependencegraph (DDG). In this paper, the real time transaction data update method based onDDG that inside vehicle is proposed. Then, considering the influence to the datavalidity and system workload by data update period, the real time data update periodis adjusted. The contributions in this paper are as follows:(1) The analytic hierarchy process (AHP) is improved for a new real time dataupdate algorithm design in vehicle control system database. When limited by thetransaction’s deadline, the data of the DDG in the real time transaction can not be allupdated. The impacts of the data to derived data are considered by weight for datasorting in DDG in order to select the relative important data. The weight of the datacan be well calculated by AHP, and the obtained update order is optimal. The accuracyof derived data is high, and the transaction deadline miss ratio is reduced.(2) The data update period is dynamic adjusted based on data similarity andmobile network disconnection probability, thus the updated data accuracy ismaintained and the system load is reduced for mobile real time transactions. In VANET communication environment, when the road conditions are stable, theupdated data in continues time intervals have very high similarity. Thus, the updateperiod can be increased for system load reducing. Otherwise, the update period shouldbe decreased for data accuracy. The mobile network disconnection probability reflectsthe difficulty for data acquisition from other vehicles. So the disconnectionprobability is used for similarity interval revised to ensure real time data updating assoon as possible in severe communication environment.