Research On Adaptive Gearshift Decision Strategy Base On Driver-Vehicle-Environment Recognition

Moving vehicle is a complex driver-vehicle-environment closed-loop system.However, the traditional shift law only works in fixed mode, which cannot satisfy thecustomers’ improving demands for vehicle performance. Actually, with the developmentof intelligent control technology, gearshift decision strategy has entered intelligent andadaptive stage. The first problem of adaptive gearshift decision method is how torecognize driver-vehicle-environment system, and the next problem is how to gearshiftdecision-making based on the recognition results. The paper focused on the several keyissues on the adaptive gearshift decision field and carried out the following researches:1) Theoretical basis of shift law. To provide the theoretical foundation for adaptivegearshift decision strategy, the simulation platform of powertrain system of automaticvehicles was established. Then the basic theory, the formulating method as well as theadaptability of the shift law were analyzed. The analysis results showed that thetwo-parameter shift law is based on the steady-state model of the engine and vehicle,which can only achieve the optimal performance for a designed working condition;however, the dynamic three-parameter power shift law has the ability to adapt to theenvironment because the vehicle acceleration is considered.2) Quantization method and quantitative analysis of driving style. By analyzing theessential relationship between driving style and driver power demand, the method ofquantifying driving style based on the driver power demand was proposed. Then thedriving style quantization experiment and the corresponding quantitative analysis workwere carried out. The quantization method is able to achieve the comprehensivequantification of the driving style and the separate quantification of the subjective and objective factors influencing driving style. The quantitative results can be used torecognize the driving style.3) Driving style recognition and corresponding adaptive gearshift decision method.The concepts of long-term and short-term driving style were defined. The former reflectedthe overall and periodic trends of driving style, and the latter reflected the driver’sinstantaneous driving intention. The method of predicting long-term driving style based onthe exponential smoothing method, classifying short-term driving style based onclustering analysis and extracting fuzzy rules of short-term driving style recognitionautomatically based on the clustering results were proposed. At last, the recognition resultsof two driving styles were merged together into the driver power demand factor thatreflected the driver’s power requirement to vehicle. The gearshift decision method basedon the driver power demand factor was proposed to realize the adaption to the driver’slong-tern driving habit and instantaneous driving intent.4) General driving environment recognition based on load degree and correspondingadaptive gearshift decision method. The concept of load degree was defined, whichsynthetically reflected the traveling resistance produced by slope, load, weather conditionand road condition. Then, the load degree based driving environment recognition methodwas proposed. In fact, the load degree essentially reflects the power demand of drivingenvironment. And then, the positive and negative load degree based gearshift decisionmethod were proposed, which were applied to conditions such as uphill, large load, anddownhill. Compared with dynamic three-parameter power shift law, gearshift decisionmethods based on load degree also has the adaptive ability to general driving environment.But its presentation was concise and is easier to implement, so it can be easily used in thereal vehicle control.5) Special driving environment recognition and corresponding gearshift decisionmethod. By extracting the vehicle lateral acceleration and its change rate from the fourwheel speed signals, meanwhile considering the current vehicle velocity, the fuzzy methodto recognize the urgency degree of curve driving based on four-wheel speeds was proposed. And then, the gear correction strategy was formulated and the experimentalverification was carried out. Through the time-domain and frequency-domain analysis ofthe engine speed variation rate signal on suburb and mountain area, the rough road featuredetection method based on engine speed variation rate was proposed, which was appliedto prevent unexpected or busy shift problems on rough road.6) Gearshift decision method on braking condition. Based on the analysis of thenecessity of engine assisted braking, shift decision method to improve braking effect anddriving safety based on fuzzy reasoning was proposed. The fuzzy reasoning rulesconsidered braking time, braking deceleration, load degree of driving environment andvehicle velocity were set up and road test was carried out. The results showed that thisgearshift decision method on braking condition can enhance the braking effect, extend thelife of the brake system, and improve the driving safety.7) Vehicle test. On the basis of vehicle experimental platform, a series of vehicle testsfor checking and verifying the driver-vehicle-environment recognition methods andcorresponding adaptive gearshift decision strategy were carried out. The results showedthat the driver and the driving environment were identified effectively without anyadditional sensors, the intelligent gearshift decision strategy based on these recognitionresults were realized, and these methods have more practical application values.