【正文】 m is fixed unless the manufacturer updates it with costly replacements. Finally, because of the limited puting resources of a vehicle (slow processor and less information storage space), it’s difficult to do much more than limit checking type diagnostics. Advanced signal analysis techniques such as signal transformations or machine learning techniques are not available. With the rapid development of the CPU and signal processing, offboard diagnostic techniques are more promising than onboard diagnostics. Standards are in place (ISO 9141) for the data link from the onboard puter to the offboard unit, so data can be collected from the ECU and analyzed offline by powerful puters. Unfortunately, at this time, diagnostic techniques lag far behind data collection techniques.Vehicle diagnosis techniques can be divided into two classes: modelbased and modelfree. Modelbased techniques employ mathematical models of the dynamics of the vehicle ponents to analyze the behavior of vehicle systems [2], [10], [12]. While these models may useful for examining simplified versions of each of the engine ponents, we do not have accurate models for a real vehicle with many interactive ponents. Modelfree systems are knowledgebased, incorporating professional knowledge from engineers without exact information regarding the details of system dynamics. The rationale behind this approach is that many experienced technicians can find faults even though they do not have extensive knowledge of the mechanical or electrical dynamics of the vehicle. Examples of such system include Strategy Engine (HP), TestBench (Carnegie Group), IDEA (Fiat Research Center), and MDS (DaimlerBenz Research) [25].. In multiple signal systems, changes in one signal generally result in changes in one or more other signals. In (a) signal B is the smoothed version of signal A, a simple relationship. In (b) each edge is indicating the feature the signal at the tail end is causing in the signal at the head end. Notice that each signal effects and is effected by multiple different signals.In this paper, we describe an offboard and modelfree diagnostic system for identifying faulty vehicle behavior through analysis of ECU signals. The signals discussed here e from the Power train Control Module (PCM) of the ECU, however, the methods developed are sufficiently general to allow for use in other multisignal fault diagnosis problems.In a tightly coupled system such as a vehicle power train, inputs and outputs from every ponent effect most of the other ponents of the system. For example, the driver pressing the throttle causes an increase in the airflow to the engine. The PCM changes the control strategy modifying fuel delivery and spark timing. Increased fuel and air increases RPM, which, in turn, dramatically changes the behavior of the transmission and other ponents such as alternator output. Furthermore, there are feedback loops in the system. The onboard controller monitors output exhaust quality, gear changes and airflow changes further modifying system behavior to keep performance at a maximum and exhaust pollution at a minimum. Outside factors such as road quality, road gradient, vehicle weight, active accessories, etc. provide physical feedback to the system further altering behavior. In our system, we capture these physical events and dependencies through the power train signals. For example, Fig. 1(a) demonstrates a simplification of the relationship between the throttle position (TP) and revolution per minute (RPM) signals. TP makes a sudden rise and fall while RPM mimics this behavior but more smoothly. This simplification is not pletely accurate but demonstrates the key point that important physical relationships can be seen through the vehicle signals. Fig. 1(b) shows a more typical set of relationships between four different signals. Each circle is a signal and each edge indicates a feature that the tail signal influences in the head signal. These relationships are often plex, include five to ten different important signals, and have many cyclic dependencies between signals. We note several important issues related to using signals to diagnose a vehicle. First, we must differentiate between a bad signal and bad vehicle behavior reflected in the signal. A bad signal is generally caused by a bad PCM or a bad sensor. Bad vehicle behavior can be caused by any of a number of factors, physical or electronic. Our system detects signal features that indicate bad vehicle behavior, whether it is caused by bad electronic parts or physical faults. Second, we note that not all of the physical dependencies present in the actual vehicle can be modeled with corresponding signals. For instance, there is no signal to indicate road bumpiness, a physical factor that can effect vehicle and, therefore, signal behavior. To handle these unknown conditions we train with vehicle data in several conditions while avoiding extreme driving conditions (., offroad racing). Finally, the same signals are not available from all vehicles. When considering behavior that depends on signal relationships, this can lead to an inability diagnose certain faults that depend on information present in the missing signal.In this paper, we focus on developing techniques of deposing multiple signals, diagnostic feature extraction, and intelligent diagnosis. The paper is organized as follows. In Section II we briefly introduce the diagnostic system. In Section III, an automatic segmentation algorithm based on wavelet multiresolution analysis is introduced. In Section IV we discuss how we can process and bine feature and segment information to form feature vectors suitable for input into a machine learning system. Section V describes how a fuzzybased machine learning system can be used to learn good and bad signal behavior. Section VI describes the implemented diagnostic system and the encouraging experimental results we’ve obtained. Finally, S