SYSTEMS AND METHODS FOR PARTICLE FILTER REGENERATION

Abstract

Method comprising: at the start of a driving cycle, displaying, by an electronic control, a first driving distance in response to each of a particulate filter (PF) loading and a previous driving history; and while driving along the first route, displaying, by the electronic control, an updated route in response to each of traffic conditions and a comparison of a real-time driving history along the first route on the drive cycle relative to the previous driving history.

Description

Area

[0001] This description generally relates to methods and systems for selecting a route for a vehicle for particulate filter regeneration based on driver behavior. General state of the art / brief description

[0002] Emissions control devices, such as particulate filters (PFs), can reduce the level of soot emissions from an internal combustion engine by capturing soot particles. Such devices can be passively regenerated during engine operation to reduce the amount of particulate matter captured. However, during vehicle operation, conditions for continuous full regeneration of the PF may not be available. For example, during urban driving conditions that include frequent idle stops and light-load engine operation, frequent premature regeneration terminations may occur. Premature regeneration terminations may occur due to driving behavior of the vehicle driver (also referred to herein as the operator), such as frequent brake application. These premature terminations may trigger the need for active regeneration, resulting in increased regeneration fuel penalty.

[0003] Various approaches for regenerating a particulate filter during a vehicle drive cycle are provided. In one example, as shown in US Pat. No. 8,424,294 B2, Schumacher et al. disclose a method for controlling the regeneration cycles of an exhaust aftertreatment system, such as a particulate filter, based on driver-specific information so that optimal regeneration is achieved. The driver-specific information may include driving habits, drive cycles, and driving routes used by the driver. Such driver-specific information may be used to predict phases during driving when regeneration of the particulate filter may be possible.

[0004] However, the inventors of the present invention have recognized potential drawbacks associated with the aforementioned approach. For example, driver-specific information may not be constant throughout a drive cycle, resulting in a significant difference between predicted driver behavior along a route and real-time driver behavior. As a result, a route planned based on predicted driver behavior without considering temporal changes in driver behavior may exhibit a PF regeneration efficiency that differs from the actual PF regeneration efficiency. Furthermore, environmental factors, including traffic and weather conditions, can significantly impact the possibility of regeneration completion over the drive cycle.Furthermore, selecting a route based solely on driver preferences may result in higher fuel consumption and increased driving time.

[0005] In one example, the problems described above may be addressed by an engine method comprising: at the start of a drive cycle, displaying a first driving route in response to each of a particulate filter (PF) loading and a previous driving history; and while driving along the first driving route, displaying an updated route in response to each of traffic conditions and a comparison of a driving history along the first route on the drive cycle relative to the previous driving history. In this way, PF regeneration efficiency may be improved by estimating a driver's mood in real time and quantitatively using the driver's mood to recommend routes to a vehicle driver.

[0006] As one example, a vehicle controller may develop a route database for a vehicle driver based on the frequently used routes along with the driving history on each route. Each time a trip is completed, the database may be updated with information related to driver characteristics, including driving practices such as pedal input, brake usage, lane change frequency, vehicle start-stop frequency, etc. At the beginning of a drive cycle, an initial driver mood may be predicted based on the driving history (driver characteristics) as retrieved from the database. Thus, multiple driver moods may exist, and there may be a change in mood throughout the drive cycle based on factors such as traffic and weather conditions.Each mood state may correspond to a different PF regeneration factor, which may directly impact the ability to achieve a desired PF regeneration over a particular route. In response to an indication of a known target (based on driver input) or a predicted target (based on driving history and route prediction algorithms), and further based on the current PF soot level and the driver's initial mood state, one or more routes may be selected from the database and presented hierarchically to the vehicle driver. For a particular target, when the PF load is higher than a threshold and the driver is in a first mood state, a first route may provide higher PF regeneration efficiency, while a second route may provide lower PF regeneration efficiency.However, for the same destination and with a higher than threshold PF load but a different driver state of mind, the first route may have a lower PF regeneration efficiency, while the second route may have a higher PF regeneration efficiency. Navigation instructions can then be provided based on the driver selection. During the drive cycle, the driver state of mind can be updated in real time based on driver interactions with traffic and environmental conditions, such as weather. A non-homogeneous state transformation matrix can be used to determine changes in the driver's state of mind during driving. If the driver's state of mind changes, the regeneration efficiencies of the routes can be recalibrated, and an alternative route, now providing the highest PF regeneration efficiency, can be displayed.The ranking of the selected tracks can be adjusted in real time based on the driver's current state of mind, so that the track at the top of the list can correspond to the highest level of achievable PF regeneration.

[0007] By taking into account the driver's current mood when selecting and ranking routes for particulate filter regeneration, the likelihood that a driver will follow the recommended route is increased. By estimating the driver's mood in real time based on driver interactions with traffic and environmental conditions and updating the ranking of the displayed routes, the probability of achieving a desired level of PF regeneration can be improved. By maintaining a database of frequently driven routes with information including the actual level of PF regeneration achieved on each route and the driver's history on those routes, it may be possible to select one or more routes from a database based on the PF regeneration requirement during a future drive cycle.The technical effect of using a non-homogeneous transition matrix to estimate changes in driver mood during the driving cycle is that the continuous inclusion of traffic scenarios can be optimally captured while determining the current driver mood and the impact on PF regeneration. By correlating each individual mood with a regeneration factor, the influence of driver behavior on regeneration can be quantified and accounted for during route planning and passive PF regeneration. By estimating suitable routes for PF regeneration while considering the influence of the driver mood, the system regeneration can be performed opportunistically, thereby reducing soot overload in the particulate filter and improving engine performance and particulate filter health.

[0008] It should be understood that the foregoing summary is provided to introduce, in a simplified form, a selection of concepts further described in the detailed description. It is not intended to identify important or essential features of the claimed subject matter, the scope of which is defined solely by the claims following the detailed description. Furthermore, the claimed subject matter is not limited to implementations that address any disadvantages noted above or in any part of this disclosure. Short description of the drawings Fig. 1 shows an embodiment of an engine system which includes a particulate filter. Fig. 2 shows a flowchart illustrating an example method that may be implemented to select a route based on PF regeneration requests. Fig. 3 shows a flowchart illustrating an example method that may be implemented to determine a current mood of a driver and the influence of the driver's mood on route selection. Fig. Figure 4 shows a flowchart illustrating an exemplary method that can be implemented to update a database of frequently traveled routes. Fig. 5A shows a first example display of suggested routes based on PF regeneration requests and an initial driver state of mind. Fig. 5B shows a second example display of suggested routes based on PF regeneration requests and an updated driver state of mind. Fig. Figure 6 shows a state machine diagram for the transition of the driver's mood. Fig. Figure 7 shows a transition matrix for changes in the driver’s mood. Fig. Figure 8 shows a table of regeneration influencing factors corresponding to each driver’s mood. Fig. Figure 9 shows a table of scaling factors corresponding to each regeneration influence factor for weights of cost functions assigned to a route. Fig. Figure 10 shows an example of the prediction and dynamic selection of a proposed route for PF regeneration. Detailed description

[0009] The following description relates to systems and methods for selecting a route from a database containing details of frequently driven routes and driver behavior for optimal particulate filter regeneration. An exemplary engine system including a particulate filter is described in Fig. 1. An engine controller may be configured to perform control routines, such as the exemplary routine of Fig. 2 to select a route from a database based on PF regeneration requests. A control routine, such as the example in Fig. 3, may be performed to estimate the current state of mind of the driver and further determine the influence of the current state of mind on the selection of driving routes that are optimal for PF regeneration. After each drive cycle, the controller may perform a routine, such as the example routine from Fig. 4 to update the database with information learned during the drive cycle. Fig. 5A and Fig. 5B show example displays of suggested routes based on PF regeneration requirements and a driver's state of mind. As in the state machine diagram of Fig. As shown in Figure 6, the driver’s state of mind can change between several states during a driving cycle and such changes can be estimated based on a transition matrix as shown in Fig. 7. The individual moods of the driver can correspond to different regeneration influencing factors, as shown in Fig. 8, and these regeneration influencing factors can influence the weightings of cost functions for route determination, as shown in Fig. 9. A prophetic example, which includes the prediction and selection of suggested routes based on PF regeneration requirements, is shown in Fig. 10 shown.

[0010] Fig. 1 schematically illustrates aspects of a vehicle system 102 having an exemplary engine system 100 including an engine 10. In one example, the engine system 100 may be a diesel engine system. In one example, the engine system 100 may be a gasoline engine system. In the illustrated embodiment, the engine 10 is a boosted engine coupled to a turbocharger 113 including a compressor 114 driven by a turbine 116. Specifically, fresh air is introduced into the engine 10 along the intake passage 42, via the air cleaner 112, and flows to the compressor 114. The compressor may be any suitable intake air compressor, such as an engine-driven or driveshaft-driven supercharger compressor.In the engine system 10, the compressor is a turbocharger compressor that is mechanically coupled to the turbine 116 via a shaft 19, the turbine 116 being driven by expanding engine exhaust gases.

[0011] As in Fig. 1, the compressor 114 is coupled to the throttle valve 20 through the charge-air cooler (CAC) 17. The throttle valve 20 is coupled to an engine intake manifold 22. From the compressor, the compressed air charge flows through the charge-air cooler 17 and the throttle valve to the intake manifold. Fig. 1, the pressure of the air charge within the intake manifold is sensed by a manifold air pressure (MAP) sensor 124.

[0012] One or more sensors may be coupled to an inlet of compressor 114. For example, a temperature sensor 65 may be coupled to the inlet for estimating a compressor inlet temperature, and a pressure sensor 66 may be coupled to the inlet for estimating a compressor inlet pressure. As another example, a humidity sensor 67 may be coupled to the inlet for estimating a humidity of an air charge entering the compressor. Still other sensors may include, for example, air-fuel ratio sensors, etc. In other examples, one or more of the compressor inlet conditions (such as humidity, temperature, pressure, etc.) may be inferred based on engine operating conditions.Additionally, when exhaust gas recirculation (EGR) is enabled, the sensors can estimate temperature, pressure, humidity, and air-fuel ratio of the air-charge mixture, including fresh air, recirculated compressed air, and residual exhaust gases ingested at the compressor inlet.

[0013] A wastegate actuator 92 may be actuated to an open position to vent at least a portion of the exhaust pressure from upstream of the turbine via the wastegate 90 to a location downstream of the turbine. By reducing the exhaust pressure upstream of the turbine, the turbine speed may be reduced, helping to reduce compressor surge.

[0014] The intake manifold 22 is connected to a series of combustion chambers 30 through a series of intake valves (not shown). The combustion chambers are further coupled to the exhaust manifold 36 via a series of exhaust valves (not shown). In the depicted embodiment, a single exhaust manifold 36 is shown. However, in other embodiments, the exhaust manifold may include a plurality of exhaust manifold sections. Configurations including a plurality of exhaust manifold sections may allow wastewater from different combustion chambers to be routed to different locations in the engine system.

[0015] In one embodiment, each of the exhaust and intake valves may be electronically actuated or controlled. In another embodiment, each of the exhaust and intake valves may be cam-actuated or controlled. Regardless of whether electronic actuation or cam-actuation is used, the timing of the opening and closing of the exhaust and intake valves can be adjusted as required for the desired combustion and emissions control performance.

[0016] One or more fuels, such as gasoline, alcohol-fuel blends, diesel, biodiesel, compressed natural gas, etc., can be supplied to the combustion chambers 30 via an injector 69. The fuel can be supplied to the combustion chambers via direct injection, port injection, throttle body injection, or any combination thereof. Combustion in the combustion chambers can be initiated via spark ignition and / or compression ignition.

[0017] As in Fig. 1, exhaust gas is directed from the one or more exhaust manifold sections to the turbine to drive the turbine 116. The combined flow from the turbine and wastegate may then flow through the exhaust aftertreatment devices 170 and 172. In one example, the first exhaust aftertreatment devices 170 may be a light-off catalyst, and the second exhaust aftertreatment devices 172 may be a particulate filter, such as a regenerable particulate filter (PF). As one example, the PF may be a diesel particulate filter coupled to the exhaust passage 104 of a diesel engine. In another example, the PF may be a gasoline particulate filter coupled to the exhaust passage 104 of a gasoline engine. The PF may be made from a variety of materials, including cordierite, silicon carbide, and other high-temperature oxide ceramics. Thus, the PF has a finite capacity to capture soot.Therefore, it may be necessary to regenerate the PF regularly to reduce soot deposits in the filter so that the flow resistance due to soot accumulation does not reduce engine performance. Passive PF regeneration can be advantageously performed during certain engine operating conditions, such as during higher engine load, when exhaust gas flowing through the PF consists of a defined composition and is above a threshold temperature to combust or oxidize the captured particulate matter. During passive PF regeneration, soot can be opportunistically combusted due to the higher exhaust temperature and also the presence of a desired amount of oxygen in the exhaust.Filter regeneration can be achieved by actively heating the filter by passing an electric current to a temperature where soot particles are burned at a faster rate than new soot particles are deposited, for example, at 400–600°C, as during active PF regeneration. During active regeneration, ignition timing can be retarded or fuel enrichment can be performed to increase exhaust gas temperature. For this reason, active PF regeneration can increase fuel consumption and parasitic energy loss (by supplying electricity to the filter). In comparison, actively heating the PF using electric current, ignition retardation, and / or fuel enrichment during passive regeneration may not be desirable.In one example, the PF may be a catalyzed particulate filter containing a washcoat of precious metal, such as platinum, to reduce the soot combustion temperature and also to oxidize hydrocarbons and carbon monoxide to carbon dioxide and water.

[0018] In one example, the exhaust aftertreatment device 170 may be configured to NO x from the exhaust stream when the exhaust stream is lean, and the captured NO x to reduce when the exhaust flow is rich. In another example, the exhaust aftertreatment device 170 may be configured to reduce the NO x -ratio to change or NO xusing a reducing agent. In yet another example, exhaust aftertreatment device 170 may be configured to oxidize hydrocarbon and / or carbon monoxide residues in the exhaust stream. Various exhaust aftertreatment catalysts with such functionality may be disposed in washcoats or elsewhere in the exhaust aftertreatment stages, either separately or together. The treated exhaust from exhaust aftertreatment devices 170 and 172 may be released, in whole or in part, to the atmosphere via main exhaust passage 104 after passing through a muffler 174.

[0019] An exhaust gas recirculation (EGR) supply passage 180 may be coupled to the exhaust passage 104 downstream of the turbine 116 to provide low-pressure EGR (LP EGR) to the engine intake manifold upstream of the compressor 114. An EGR valve 62 may be coupled to the EGR passage 180 at the junction of the EGR passage 180 and the intake passage 42. The EGR valve 62 may be opened to admit a controlled amount of exhaust gas to the compressor inlet for a desired combustion and emissions control performance. The EGR valve 62 may be configured as a continuously variable valve or as an on / off valve. In further embodiments, the engine system may include a high pressure EGR flowpath wherein exhaust gas is drawn from upstream of the turbine 116 and recirculated to the engine intake manifold downstream of the compressor 114.

[0020] One or more sensors may be coupled to the EGR passage 180 to provide details regarding the composition and conditions of the EGR. For example, a temperature sensor may be provided to determine a temperature of the EGR, a pressure sensor may be provided to determine a pressure of the EGR, a humidity sensor may be provided to determine a humidity or water content of the EGR, and an air-fuel ratio sensor may be provided to estimate an air-fuel ratio of the EGR. Alternatively, EGR conditions may be inferred by the one or more temperature, pressure, humidity, and air-fuel ratio sensors 65-67 coupled to the compressor inlet. In one example, the air-fuel ratio sensor 57 is an oxygen sensor.

[0021] A variety of sensors, including an exhaust temperature sensor 128 and an exhaust oxygen sensor 129 and an exhaust pressure sensor 129, may be coupled to the main exhaust passage 104. The oxygen sensor may be a linear oxygen sensor or UEGO (universal or wide-range exhaust gas oxygen) sensor, a dual-state oxygen sensor, or an EGO, HEGO (heated EGO), NOx, HC, or CO sensor.

[0022] The engine system 100 may further include the control system 14. The control system 14 is shown receiving information from a plurality of sensors 16 (various examples of which are described herein) and sending control signals to a plurality of actuators 18 (various examples of which are described herein). A navigation system 154, such as a global positioning system (GPS), may be coupled to the control system 14 to determine the location of the vehicle 102 at a key-on event and at any other time. The navigation system may be connected to an external server and / or a network cloud 160 via wireless communication 150. The navigation system 154 may determine the current location of the vehicle 102 and obtain environmental condition data (such as temperature, pressure, etc.) and road information (such as road gradient) from a network cloud 160.The controller 12 may be coupled to a wireless communication device 152 for direct communication of the vehicle 102 with a network cloud 160. Upon completion of a drive cycle, the database 13 may be updated with route segment information, including driver behavior, driver mood states, a level of particulate filter regeneration achieved, engine operating conditions, date and time information, and traffic information. Further, details of the trip, including the start and destination, stops during the trip and the duration of each stop, the road gradient (terrain) of the route, fuel consumption, duration of the trip, driver behavior, etc., may be stored in a database 13 within the controller 12. Information regarding the route and traffic may be learned from the navigation system 154 and the external cloud 160 via wireless communication 150.The various routes in the database can be compared and ranked based on fuel efficiency, trip duration, and achievable PF regeneration level. Details regarding a driver's driving pattern can be retrieved from the controller's memory and used to rank the routes. Furthermore, the vehicle driver's driving pattern can be learned over a number of vehicle drive cycles based on one or more common drive time patterns, habitual probability patterns, route-based statistical profiles, and environmental feature profiles. Other statistical profiles, different driver mood states, and conditions for transitioning from one mood state to another can be learned and stored in database 13. The engine operating parameters can be estimated using inputs from one or more sensors 16, and the information can be added to database 13.Details on updating database 13 are described in detail in relation to . Fig. 4 discussed.

[0023] At the start of a drive cycle (at the vehicle key-on event), a request for PF regeneration during the upcoming drive cycle may be assessed based on the PF soot level. In response to the driver providing a destination (such as via an input to an in-vehicle navigation system), one or more routes may be selected from the database 13 based on the driver's mood to enable a higher degree of PF regeneration while optimizing fuel efficiency and travel time. Route selection may further be based on a driver-selected cost function, including the highest fuel efficiency and the lowest travel time for the drive cycle.The one or more selected routes may then be ranked as a weighted function of each of a particulate filter regeneration efficiency, a probability of completing a PF regeneration event, a fuel efficiency, and a travel time of each of the one or more routes. For example, if based on the current PF soot level, it is inferred that PF regeneration is desired during the upcoming drive cycle, where the highest-ranked route displayed to the vehicle operator may be a route that allows the target to be achieved while providing the highest level of regeneration and while providing a certain level of fuel efficiency.In situations where the driver does not select a route from the one or more routes recommended (displayed) to the driver, an upcoming route segment may be dynamically predicted based on a driving history (of the driver) retrieved from the database. Furthermore, the driver may select a route from the one or more recommended routes and begin driving along the route and then deviate from the selected route. During such deviations, the controller may dynamically predict an upcoming route segment based on a driving history of the driver (such as preferred routes of travel during a particular time of day or day of the week) retrieved from the database.One or more routes or route segments may be selected from the database based on the predicted target, ranked for their particulate filter regeneration efficiency, probability of completing a PF regeneration event, fuel efficiency, and travel time of each of the one or more routes, and displayed to the driver. Details of the route selection based on information stored in the database 13 are described with reference to [ ]. Fig. 2 discussed.

[0024] In response to a driver destination selection indicated via a display of a vehicle, a particulate filter soot load may be estimated, a current location of the vehicle may be estimated, one or more routes may be retrieved from a database from the current location to the destination, the one or more routes may be ranked based on each of a particulate filter regeneration efficiency, fuel efficiency, and travel time of each route; and the one or more routes to the selected destination may be displayed to the driver in order of ranking.

[0025] The selection and classification of one or more routes may further be based on a driver's mood. A driver's mood may represent the driver's real-time driving behavior, so the driver may drive more aggressively in one mood, while the driver may drive more relaxed in another mood during the drive cycle. At the start of a drive cycle, an initial driver's mood may be selected from a plurality of driver's moods stored in database 13 based on the driver's previous driving history, traffic conditions at the start of the drive cycle, and environmental conditions at the start of the drive cycle, including temperature, humidity, precipitation, etc.The driver's previous driving history includes routes driven by the vehicle driver depending on one or more of the time of day, the day of the week, the start and end of the driving cycle, and driving characteristics, including brake usage frequency, average acceleration force used, and average lane change frequency. The driver's initial state of mind may correspond to a first particulate filter regeneration factor. Based on the first particulate filter regeneration factor, the selection and classification of the one or more routes may be updated.Updating the ranking of the one or more routes includes ranking each of the one or more routes based on a weighted function of each of the corresponding regeneration completion efficiencies, a probability of completing a particulate filter regeneration event, a fuel efficiency, and a time-to-destination of each of the one or more routes, wherein the weighted function is ranked based on the first regeneration factor. Once a driver selects a route from the one or more ranked routes, navigation instructions for the driver-selected route may be displayed to the driver.

[0026] As the driver travels along the driver-selected route, real-time driver interactions with traffic may be learned, including one or more of stop frequency, lane change frequency, accelerator input, and brake input during the drive cycle. Based on the learned real-time driver interactions with traffic and a comparison of the real-time driver interactions with traffic while traveling along the driver-selected route relative to the previous driving history, the driver's state of mind may be updated from the first state to a second state of mind, wherein the second state of mind is also selected from the database 13. The updated second driver's state of mind may correspond to a second particulate filter regeneration factor.In response to the change in the driver's mood, the ranking of the one or more routes may be updated as the weighted function of each of the corresponding regeneration completion efficiencies, the probability of completion of the particulate filter regeneration event, the fuel efficiency, and the time to destination of each of the one or more routes, which may be scaled based on the second regeneration factor.For example, during initial ranking, the highest ranked route displayed to the vehicle driver may be a route providing the highest level of PF regeneration. However, once the ranking is updated in response to a change in the driver's mood, the previously highest ranked route may no longer provide the highest level of PF regeneration, and another route providing the highest level of PF regeneration may now be ranked first. The updated ranking of the one or more routes may then be displayed to the driver for further selection. Upon completion of the drive cycle, a level of PF regeneration achieved during the drive cycle may be learned, and the database 13 may be updated with the learned level of PF regeneration achieved for the drive cycle, the initial driver mood, and the updated driver mood.

[0027] The control system 14 may include a controller 12. The controller 12 may receive input data from various sensors 18, process the input data, and trigger various actuators 81 in response to the processed input data based on an instruction or code programmed therein according to one or more routines. For example, the sensors 16 may include the exhaust oxygen sensor located upstream of the turbine 116, the pedal position sensor, the MAP sensor 124, the exhaust temperature sensor 128, the exhaust pressure sensor 129, the oxygen sensor, the compressor inlet temperature sensor 65, the compressor inlet pressure sensor 66, and the compressor inlet humidity sensor 67. Other sensors, such as additional pressure, temperature, air-fuel ratio, and composition sensors, may be coupled to various locations in the engine system 100.The actuators 81 may include, for example, a throttle 20, an EGR valve 62, a wastegate 92, and a fuel injector 69. As one example, during the key-on event, the controller may select an optimal route for the drive cycle based on a soot level on the PF as estimated via the exhaust pressure sensor 129, based on information stored in the database and input from the navigation system 154 and the network cloud 160. The controller may then display the optimal route to the driver, and when the route is selected, a level of PF regeneration achieved during the drive may be monitored and learned.

[0028] In some examples, vehicle 102 may be a hybrid vehicle with multiple torque sources available to one or more vehicle wheels 55. In other examples, vehicle 102 is a conventional engine-only vehicle or an electric vehicle with only electric machine(s). In the example shown, vehicle 102 includes an engine 10 and an electric machine 52. Electric machine 52 may be an electric motor or an electric motor / generator. The crankshaft of engine 10 and electric machine 52 are connected to vehicle wheels 55 via a transmission 54 when one or more clutches 56 are engaged. In the example shown, a first clutch 56 is provided between crankshaft 140 and electric machine 52, and a second clutch 56 is provided between electric machine 52 and transmission 54.The controller 12 may send a signal to an actuator of each clutch 56 to engage or disengage the clutch to connect or disconnect the crankshaft to or from the electric machine 52 and its associated components and / or to connect or disconnect the electric machine 52 to or from the transmission 54 and its associated components. The transmission 54 may be a manual transmission, a planetary gear system, or another type of transmission. The powertrain may be configured in a variety of ways, including as a parallel, series, or series-parallel hybrid vehicle.

[0029] The electric machine 52 receives electrical power from a traction battery 58 to provide torque to the vehicle wheels 55. The electric machine 52 can also be operated as a generator, for example, to provide electrical power to charge the battery 58 during braking.

[0030] In this way, the components from Fig. 1 a vehicle system comprising: a vehicle, a navigation system wirelessly connected to an external network, a display, an engine including an intake system and an exhaust system, the exhaust system including a particulate filter (PF) coupled to an exhaust passage and a pressure sensor coupled to the exhaust passage upstream of the particulate filter, and a controller having computer-readable instructions stored in non-transitory memory for: at a start of a drive cycle, displaying a first route based on PF loading and a first driver mood, and in response to driver interactions with traffic while driving on the first route, displaying a plurality of updated routes based on a second driver mood,wherein the first driver state of mind is selected from a database based on each of the PF load and a driving history, and a change from the first driver state of mind to the second driver state of mind is based on the driver interactions with traffic while driving on the first route.

[0031] Fig. 2 shows an exemplary method 200 for selecting a route based on particulate filter (PF) regeneration requirements. Instructions for performing method 200 and the other methods included herein may be provided by a controller based on instructions stored in a memory of the controller and in conjunction with sensors of the engine system, such as those described above with reference to Fig. 1. The controller may use motor actuators of the motor system to adjust motor operation according to the procedures described below.

[0032] At 202, the routine may include determining whether a vehicle key-on event is detected. For example, it may be determined that the vehicle operator has expressed an intent to initiate vehicle operation. Thus, acknowledging a vehicle key-on event indicates an upcoming vehicle drive cycle. While referred to herein as a vehicle "key-on" event, it is understood that the operator may indicate an intent to operate the vehicle with or without the use of a key. For example, vehicle operation may be initiated by inserting a key (active key) into an ignition slot and moving the slot to an "ON" position. Alternatively, vehicle operation may be initiated when a key (passive key) is within a threshold distance of the vehicle (e.g., inside the vehicle).As another example, vehicle operation may be initiated when the driver pushes an ignition button to the "ON" position. Still other approaches may be used by a driver to indicate intent to operate the vehicle. Thus, the vehicle driver's driving patterns can only be learned when the vehicle is being operated. Therefore, if a vehicle key-on event, and thus an upcoming vehicle drive cycle, is not confirmed, the process may terminate and PF regeneration may not be performed.

[0033] When a key-on event is confirmed, current vehicle and engine operating conditions may be estimated and / or measured at 204. These may include, for example, engine rpm, vehicle speed, engine temperature, engine load, ambient conditions (such as ambient humidity, temperature, and barometric pressure), boost level, exhaust gas temperature, manifold pressure, manifold airflow, battery charge level, etc.

[0034] At 206, the level of soot collected in an exhaust PF may be determined based on an input from an exhaust pressure sensor (such as pressure sensor 129 of Fig. 1), which is positioned upstream of the PF. As the soot level in the PF increases, exhaust backpressure can increase pumping losses, thereby affecting engine performance and increasing fuel consumption. Thus, when the soot level increases above a threshold, the PF can be regenerated by burning at least some of the soot deposited on it. However, passive regeneration of the PF can be adversely affected during engine operating conditions such as idling, lower engine load, and lower engine temperature. Incomplete or aborted PF regenerations can negatively impact engine efficiency. Therefore, when starting a new trip, a trip route can be selected taking into account the soot level on the PF so that PF regeneration can be performed opportunistically during the trip.

[0035] At 208, the routine includes determining whether a destination has been set by the driver. The driver may set a destination via an input to the in-vehicle navigation system. If it is determined that the destination is known, the routine proceeds to step 210 to retrieve one or more routes between the start and the destination from a database (such as database 13 in Fig. 1). The starting point (such as coordinates, geographical location) can be determined using the vehicle's on-board navigation system or a network cloud via a wireless connection. The database is further updated with information on frequently traveled vehicle routes. Information including the starting point and destination, routes traveled, stops during the trip and the duration of each stop, traffic information for each route, the day and time of the trip, engine operating conditions, fuel consumption, duration of the trip, the possible degree of PF regeneration, driver driving characteristics, etc., can be available in the database. An exemplary method for updating the database during each trip is described with reference to Fig. 4. Given the current vehicle position and the destination (as specified by the driver), a variety of possible routes may be available in the database.

[0036] At 212, one or more routes can be selected from the database for the journey between the current vehicle position and the destination. Dynamic programming can be performed to estimate the costs associated with each route. As an example, the cost function associated with a route can be estimated using Equation 1. JA−B=∑ABw1(1−E(θ))+w2mfuelregen+w3mfuelA−B+w4tA−B+w5P(Abort|θ<θ*) where J A-B is the total cost function (sum of the individual cost functions) associated with a specific route from start A to destination B, E(θ) is the expected PF regeneration level of this route from point A to B, w1 is the weighting associated with the expected regeneration level of this route, w2mfuelregen the cost of heating the PF to a temperature at which regeneration can begin, w3mfuelA−B the fuel consumption for the journey in connection with the distance from A to B, including fuel use due to pumping losses caused by back pressure with increasing soot load on the PF, w4t A-B the time taken to travel from A to B along the given route, P(A bort |θ < θ*) is the probability of a PF regeneration process to be completed during the journey from point A to B over that particular distance, and w5 is the weight associated with the probability of completion of the PF regeneration.

[0037] Weights w1 and w5 can be adjusted based on the PF soot level. In one example, if the current PF soot level is above a threshold and PF regeneration is desired during the upcoming drive cycle, each of weights w1 and w5 can be increased and weights w3 and w4 can be decreased to increase the cost functions of the routes with the lower expected PF regeneration level or with a higher probability of completion of the PF regeneration process. However, if PF regeneration is not desired during the upcoming drive cycle, weights w1 and w5 can be decreased, while weights w3 and w4 can be increased to favor the routes with lower travel time and higher fuel efficiency.

[0038] The weights w1 and w5 can be further adjusted based on the driver's current mood. A driver's mood can influence driving characteristics, such as accelerator pedal application and release frequency, gear shift frequency, and brake application frequency, which can further influence PF regeneration. A driver's mood can correspond to a regeneration impact factor (RIF), and each RIF, in turn, can correspond to a scaling factor for each of the weights w1 and w5.For example, if the driver is in a first mood state where their driving characteristics may be optimal for PF regeneration (such as lower pedal application and release frequency, lower gear shift frequency, and lower brake application frequency), the corresponding RIF may result in an equal scaling factor for each of the weights w1 to w5, so that the effect of the driver's mood state may not change the total cost function associated with the route. As an example, the same scaling factor of 0.2 may be assigned to each of the weights w1 to w5. Since the scaling factors are the same for all weights (the sum of the scaling factors is always equal to one), it can be inferred that the driver's mood state may have no adverse effect on the total cost function of any route.In another example, if the driver is in a second mood state where their driving characteristics may adversely affect PF regeneration, the RIF corresponding to the second mood state may result in unequal scaling factors for each of the weights w1 and w5. To incorporate the adverse impact of the driver's mood state into the cost function estimation for the route, the corresponding RIF may result in a higher scaling factor for w1 (the weight associated with the expected PF regeneration level for that route) and w5 (the weight associated with the probability of PF regeneration completion) relative to the scaling factors associated with other weights (w2, w3, and w4).By assigning a higher scaling factor for w1 and w5, the individual cost functions related to each of the expected PF regeneration level for that route and the probability of PF regeneration completion can be increased relative to the individual cost functions related to other factors, such as travel time and fuel usage. In this way, the current driver's mood can be quantitatively accounted for in the estimation of the total cost function associated with each route between the start and a destination. Details regarding real-time determination of a driver's mood and the influence of the driver's mood on route selection are described in [References]. Fig. 3 described in detail.

[0039] Thus, traffic information, such as signal phase and timing (SPaT) information available from the external server or navigation system, may be considered while estimating the probability of a PF regeneration process being completed over a given drive cycle and predicting the duration of the trip. For example, a higher number of traffic stops and traffic congestion in general may increase both the probability of completing a passive PF regeneration event and the duration of the trip. The controller may determine each of the expected PF regeneration level and the probability of PF regeneration completion through a determination that directly considers the traffic situation, such as increasing the expected PF regeneration level and decreasing the probability of PF regeneration completion with a decrease in the number of traffic stops.Alternatively, the controller may determine each of the expected PF regeneration level and the probability of PF regeneration completion based on a calculation using a lookup table, where the input is the current traffic situation and the output is the expected PF regeneration level and the probability of PF regeneration completion.

[0040] Once the cost functions for the plurality of available routes between the start and destination have been estimated, the routes may be ranked based on the cost function, with the highest-ranked route corresponding to the lowest cost function. In one example, if PF regeneration is desired during the upcoming drive cycle, such as when the PF soot level is higher than a threshold, a route with the highest expected PF regeneration level and the lowest probability of completing the regeneration event may be ranked highest. The highest-ranked (recommended) route may be a route that allows the destination to be reached without significant delay while providing the highest level of regeneration and some degree of fuel efficiency.The subsequent route may provide a relatively lower level of regeneration while still providing some degree of fuel efficiency, and so on. In another example, if PF regeneration is not desired during the upcoming drive cycle, such as if the PF soot level is less than the threshold, the selected routes displayed may be ranked based on time to reach the destination and / or fuel cost, and the recommended route may be selected regardless of its ability to complete PF regeneration. Thus, the highest-ranked (recommended) route may be a route that allows the destination to be reached in the shortest amount of time or using the least amount of fuel.

[0041] Once one or more routes have been selected and ranked from the database, the selected routes may be displayed to the driver at 214 in order of their ranking. The screen and user interface of the in-vehicle navigation system may be used to display the selected routes to the driver.

[0042] At 216, the routine includes determining whether the driver has selected a route from the list of recommended (displayed) routes. If it is determined that the driver has selected one of the recommended routes, the routine proceeds to 218, where navigation instructions for the selected route are provided to the driver.

[0043] At 220, when the driver follows the selected route, in addition to the passive regeneration occurring, the controller may also schedule an active regeneration of the PF during travel from the start to the destination. The scheduling may be based on the soot level and upcoming road conditions and corresponding engine operating conditions. As an example, PF regeneration may be scheduled once the soot level increases above the threshold and the driving conditions are favorable for PF regeneration, such as when the engine load is higher than a threshold load and the engine temperature is higher than a threshold temperature. During the scheduled PF regeneration, the temperature of the exhaust gas may be increased by passing electricity through the PF to burn off the soot deposited on the PF, thereby reducing the soot load on the PF.The PF can be passively regenerated while driving when the exhaust gas temperature is higher than a threshold and consists of a desired chemical composition that facilitates the oxidation of the soot deposited on the PF.

[0044] At 222, the routine includes determining whether the driver took a detour (deviation) from the expected route. Each of the passive and active regenerations of the PF regeneration may be affected by unexpected changes in driving and traffic conditions. In one example, the detour may have a higher number of traffic signals compared to the expected route, and the frequent traffic stops may negatively impact the PF regeneration. If it is determined that a detour will not be taken, the specified PF regeneration may continue, and at 226, the routine includes determining whether the expected destination was reached. If it is determined that the expected destination was not reached, the specified PF regeneration may continue at 224.

[0045] Once it has been confirmed that the target has been met, the soot load on the PF at 228 can be estimated via the exhaust pressure sensor, and the data can be updated. By estimating the residual soot level in the PF, it is possible to estimate how much soot was burned during the regeneration process. Based on the level of soot removal during the drive cycle, it is possible to estimate the PF regeneration level achieved while driving along this route, as well as the probability that a PF regeneration process was completed while driving along this route.

[0046] At 230, the route database may be updated with information including fuel consumption while driving along that route, driving time (duration), traffic information, PF regeneration schedule, level of PF regeneration achieved, and the driver's mood states, including conditions that trigger a change in mood state. An exemplary method for updating the database after each trip is described with respect to Fig. 4 executed.

[0047] Returning to 208, if it is determined that a destination is not provided by the driver, the routine proceeds to 232 to predict a possible destination based on the driving history as stored in the database. As one example, the prediction may be performed while considering the current vehicle location, frequently traveled routes during a particular time of day and day of the week, and a driver's state of mind. Traffic conditions (such as traffic congestion) and weather conditions (such as a rain or snow forecast) near the current vehicle location may also be considered while predicting the destination. In one example, the prediction may be performed in increments while driving. The vehicle controller may divide the route into route segments and predict an expected destination for an initial route segment.The greedy algorithm can be used by taking optimal path segments to predict intermediate points on the way to a destination.

[0048] If it is determined at 222 that a detour has been taken and the vehicle is no longer traveling the expected route to the expected destination (as selected in step 216), the routine may also proceed to step 232, where a final destination or intermediate points may be predicted based on information available in the database. At 234, the controller may use stochastic dynamic programming based on the predicted destination (or upcoming intermediate point) to update the one or more routes selected to reach the predicted destination. The selection process may follow the algorithm shown in step 212 using Equation 1.Once the total cost functions for the plurality of routes between the vehicle's current location and the predicted destination have been estimated, the routes may be ranked based on the cost function, with the highest ranked route corresponding to the lowest total cost function. If PF regeneration is desired during the upcoming segment of the trip, a route with the highest expected PF regeneration level and the lowest probability of completing the regeneration event may be the highest recommended route. Once the one or more routes have been updated and ranked, the routine proceeds to step 214, at which point the selected routes may be displayed to the driver in order of their rankings.If at 216 any of the recommended routes are not accepted by the driver, at 236 the controller may predict the route based on the driver preference (corresponding to a driver state of mind), current traffic, and weather conditions at the current location and the destination. As one example, the driver may prefer to take a particular route during a sunny day on weekday mornings. In another example, intermediate route segments may be predicted based on the driving history as retrieved from the database. The controller may then set a PF regeneration event based on the predicted route segment.

[0049] Thus, in response to a destination for a drive cycle not being specified by a driver, a current location of the vehicle may be determined, a driving history of the driver may be retrieved, a destination may be predicted based on the driving history, the selection of one or more upcoming route segments may be dynamically updated based on the current location of the vehicle relative to the predicted destination, the route segments may be ranked and displayed to the driver based on each of a particulate filter regeneration efficiency, fuel efficiency, and drive time, wherein the one or more route segments to the predicted destination are displayed in order of ranking.

[0050] Fig. 3 illustrates an exemplary method 300 for real-time estimation of a driver's mood and the influence of the driver's mood on route selection for a trip. Method 300 may be part of method 200 and may be performed at step 212 of method 200. Route selection may be further based on the soot load accumulated on the particulate filter (PF).

[0051] At 302, the controller may learn a date of the trip, the time of the trip, including a time of day when the vehicle is traveling, what day of the week the vehicle is traveling, etc. The controller may learn this information from an in-vehicle navigation system (e.g., GPS device) or from a network cloud via a wireless connection. At 308, the controller may learn launch characteristics, including geographic location, current weather conditions, and traffic conditions. For example, the controller may determine launch characteristics based on information from the vehicle navigation system or the network cloud. In one example, the geographic location of the launch may include the GPS coordinates of the launch. Weather conditions may include temperature, humidity, wind speed, and precipitation (such as rain, snow, etc.).Traffic conditions may include the speed limit of the road on which the vehicle is operating, the speed of overall traffic movement, the average distance between vehicles, traffic congestion, etc. Additionally, the controller can learn the geographic location of the destination based on driver input to a navigation system.

[0052] At 310, the driving history, which includes driving characteristics of the driver, can be retrieved from the database (such as database 13 in Fig. 1). In one example, a driver may be identified by the specific key used by the driver to operate the vehicle. In another example, a driver may be identified based on the time, day, date of the trip, and the geographic location of the start. Thus, a particular driver (such as Driver 1) may operate the vehicle on weekdays at a specific time (or time window) of the day, while another driver (such as Driver 2) may operate the vehicle on weekends and during a specific time window. The driver's characteristics may include the frequency of brake use, average g-force used, average lane change frequency, etc. The driver's characteristics may vary based on the time, day, date of the trip, weather, and the geographic location of the start.In one example, the driver may drive more aggressively during a sunny day compared to their driving style during rainy weather (such as frequent acceleration, increased lane changes, higher speed). Other driver preferences, including more frequently traveled routes, stops during the trip, etc., may also be retrieved from the database. In one example, a driver may drive to a specific destination every weekday morning. If the driver leaves the start at a specific time, the driver may typically stop on the way to the destination. However, if the driver leaves the start at a later time, the driver may drive more aggressively to the same destination without stopping. Thus, the driver's time constraints may be different for different days of the week; for example, the driver may be more time constrained on weekdays than on weekends.In another example, the driver's preferences may depend on weather conditions. For example, the driver may take a different route than the frequently traveled route (while traveling from a start to a frequently traveled destination), such as when it is snowing or when snow is forecast, in order to avoid road segments that have an incline. In yet another example, the driver's preferences may depend on traffic conditions. For example, the driver may take a different route than the frequently traveled route if there is a traffic jam at the start. In yet another example, the driver may select a route based on fuel level. For example, the driver may select a shorter route if the fuel level in the tank is less than a threshold level.In yet another example, route selection may be based on the vehicle's load, for example, the number of passengers in the vehicle or whether a trailer is being towed. The database is updated with information related to the driver's characteristics and preferences.

[0053] At 312, the controller may assign an initial driver mood based on the retrieved information, including driving history (driver characteristics and preferences), the day and time of the trip, and starting characteristics (weather conditions and traffic conditions). The driver mood may directly influence the driver's behavior (driving style) during the trip, which may affect PF regeneration and the probability of regeneration completion. The moods are contextual and probabilistic, and each driver may exhibit a variety of moods (S DK , K = 1,2,3,....,n) and each state can have a different effect on PF regeneration. In one example, a driver can have three different states of mind, a first state S D0 , a second state S D1 , and a third state S D2 . The first state of mind (SD0 ) may correspond to an optimal state for PF regeneration. When the driver operates in this optimal state of mind, driving characteristics (such as frequency of brake use, average g-force used, average lane change frequency, etc.) may facilitate PF regeneration and not increase the likelihood of regeneration termination. For example, in the optimal state, the driver may operate the vehicle at a constant speed for longer periods while applying the brakes less frequently, the driver may not accelerate or decelerate within a short period of time, and may not change lanes frequently. In this optimal state of mind, passive regeneration of the PF can be performed at a higher level, allowing for more complete cleaning of the PF. The second driver state of mind (S D1) may correspond to a suboptimal state for PF regeneration, where the achieved regeneration level is lower than the regeneration level achieved in the first state. When the driver is driving in this suboptimal state, the driving characteristics (such as frequent acceleration and deceleration, lane changes, stopping) may reduce the PF regeneration level and / or increase the likelihood of early regeneration termination. The third driver state (S D2) may correspond to a least optimal state for PF regeneration, where the achieved regeneration level is less than the regeneration level achieved in each of the first state and the second state. When the driver is driving in this least optimal state, driving characteristics (such as frequent braking, driving at a lower speed) may further reduce the PF regeneration level and / or further increase the likelihood of early regeneration termination. In this least optimal state, passive regeneration of the PF may be frequently interrupted, and regeneration may not be performed at a desired level.

[0054] Because the driver's mood affects PF regeneration, a regeneration influence factor may be correlated with each driver's mood. At 314, an initial regeneration influence factor corresponding to the initial mood may be determined. In one example, the controller may use a lookup table to determine the initial regeneration influence factor corresponding to the initial driver's mood, where the input is the initial driver's mood and the output is the regeneration influence factor. Fig. Figure 8 shows an exemplary table 800 of the regeneration influencing factors corresponding to each mood state. The first row 802 shows a first regeneration influencing factor f0 corresponding to the optimal mood state (S D0 ). The second line 804 shows a second regeneration influence factor f1, which corresponds to the suboptimal state of mind (S D1). The third line 806 shows a third regeneration influence factor f2, which corresponds to the least optimal state of mind (S D2 ) corresponds.

[0055] At 316, a first set of scaling factors corresponding to the initial regeneration impact factor may be determined. Each regeneration impact factor may have a corresponding set of scaling factors that may be applied to weights used for the cost function calculation for each route between a start and a destination. As in step 212 of Fig. As shown in Figure 2, one or more routes may be selected from the database for travel between the current vehicle position and the destination based on a PF regeneration request. Dynamic programming may be performed to estimate the total cost associated with each of the one or more selected routes. As an example, the total cost function associated with a route may be estimated by Equation 1.The total cost function associated with a particular route between a start and a destination may be a summation of the individual cost functions corresponding to each of an expected PF regeneration level of that route from the start to the destination, a probability of a PF regeneration process to be completed during the drive cycle, the cost of warming the PF to a temperature at which regeneration can begin, fuel consumption, and the duration of the trip between the start and the destination. Weights corresponding to individual cost functions may be adjusted based on scaling factors corresponding to a regeneration impact factor. In one example, a scaling factor may be multiplied by the corresponding weight of a cost function to determine a scaled weight.

[0056] Fig. 9 shows a table 900 of scaling factors corresponding to the regeneration influence factors and the cost function weights. The first column 902 of the table 900 lists weights associated with a unit cost function for a route between a start and a destination. The second column 904 lists the unit cost components of the total cost function. The first weight w1 can be associated with the expected PF regeneration level of a route, the second weight w2 can be associated with the cost of heating the PF to a temperature at which regeneration can begin, the third weight w3 can be associated with fuel consumption, the fourth weight w4 can be associated with time duration, and the fifth weight w5 can be associated with the probability of a PF regeneration process being completed while traveling on the route.

[0057] The third column 906 of the table shows the first set of scaling factors corresponding to the first regeneration influence factor (f0) for each weighting. For the first regeneration influence factor (f0), equal scaling factors of 0.2 can be assigned to each of the weightings w1 to w5. Since the scaling factors are the same for all weightings and the sum of the scaling factors is kept at 1, it can be deduced that the first regeneration influence factor (f0) may not have a significant impact on the individual cost functions and the total cost function of any given route. In other words, for the first regeneration influence factor (f0), the driver may not implement any specific trip criteria as critical, and therefore the costs associated with all trip parameters can be weighted equally.

[0058] The fourth column 908 of the table shows the second set of scaling factors corresponding to the second regeneration impact factor (f1) for each weighting. For the second regeneration impact factor (f1), the scaling factors may be unevenly distributed while maintaining the sum of the scaling factors at 1. The scaling factors associated with each of w1 (PF regeneration level), w2 (cost of regeneration), and w5 (probability of regeneration completion) may be 0.25, while the scaling factors associated with each of w3 (fuel consumption) and w4 (trip duration) may be 0.125. By assigning a higher scaling factor to each of w1, w2, and w5, the unit costs associated with the route's PF regeneration efficiency may be increased relative to the unit cost functions related to other factors, such as trip time and fuel usage.Based on the increase in the cost functions in conjunction with the PF regeneration efficiency of the line, it can be deduced that the second regeneration influencing factor (f. a ) may negatively impact the PF regeneration efficiency for the track when attempting regeneration.

[0059] The fifth column 910 of the table shows the third set of scaling factors corresponding to the third regeneration influence factor (f2) for each weighting. For the third regeneration influence factor (f2), the scaling factors may be unevenly distributed while maintaining the sum of the scaling factors at 1. The scaling factor associated with w1 (PF regeneration level) may be 0.35, the scaling factor associated with w2 (regeneration cost) may be 0.15, and the scaling factor associated with w5 (probability of regeneration completion) may be 0.5, while the scaling factors associated with each of w3 (fuel consumption) and w4 (trip duration) may be 0.By assigning a zero scaling factor to the trip duration and fuel consumption and increasing the scaling factors for w1, w2, and w5, the individual cost functions associated with the route's PF regeneration efficiency can be further increased, while the individual cost functions related to trip time and fuel usage may not be considered in the total cost function estimation. The third regeneration impact factor (f2) can significantly influence the PF regeneration efficiency for the route because the probability of regeneration completion is higher and the total cost function for the route is calculated entirely based on the PF regeneration efficiency, thereby increasing the trip cost and attempting regeneration under these conditions.

[0060] At 318, the weights scaled by the first set of scaling factors corresponding to the initial regeneration impact factor may be used to estimate a total cost function associated with each leg between the start and a destination, as selected from the database. Estimation of a total cost function based on individual cost functions associated with the expected PF regeneration level of a leg, the probability of a PF regeneration process to be completed during the drive cycle, the cost of heating the PF to the regeneration temperature, fuel consumption, and the duration of the trip may be performed using Equation 1. A plurality of legs between the start and the destination may be selected from the database, and the total cost function may be estimated for each of the plurality of legs.The details of the total cost function estimation are given in step 212 of . Fig. 2 described.

[0061] Once the total cost functions for the plurality of available routes between the start and destination have been estimated, the routes may be ranked based on the total cost function, with the highest ranked route corresponding to the lowest total cost function. In one example, if PF regeneration is desired during the upcoming drive cycle, such as when the PF soot level is higher than a threshold, a route with the highest expected PF regeneration level and the lowest probability of completing the regeneration event may be ranked highest. The highest ranked (recommended) route may be a route that allows the destination to be reached without significant delay while providing the highest level of regeneration and some degree of fuel efficiency.The subsequent route may provide a relatively lower level of regeneration while still providing a certain degree of fuel efficiency, and so on. Once one or more routes have been selected and ranked from the database, the selected routes can be displayed to the driver in order of their ranking. The in-vehicle navigation system's screen and user interface can be used to display the selected routes to the driver.

[0062] With brief reference to Fig. 5A shows a screenshot 500 of an exemplary in-vehicle navigation system displaying the ranked route options. In this example, PF regeneration is desired during the current drive cycle, and therefore, a route with the highest expected PF regeneration level and the lowest probability of completion of the regeneration event may be ranked first. In this example, the driver's state of mind may also be the optimal, first state (S D0) and therefore there may not be a significant (adverse) influence of the driver's state of mind on the route rankings. Four routes were selected from the database and ranked in order of their expected PF regeneration level in rows 502, 504, 506, and 508, respectively. In each column, a first box 501 indicates the suggested route between the start and the finish, a second box 503 shows a PF regeneration percentage achievable while driving, and a third box 505 shows the time required to reach the finish from the start. The driver can select one of the four routes via a fourth box included in each row. The driver can select the fourth box from any of the four routes via a user interface, such as a touch function on the screen.

[0063] The first route, as shown in line 502, corresponds to the highest expected PF regeneration level, but the driving duration is the longest. The fourth route, as shown in line 508, corresponds to the lowest expected PF regeneration level, but the driving duration is the shortest. Each of the second route, as shown in line 504, and the third route, as shown in line 506, corresponds to intermediate levels of the expected PF regeneration level and driving duration. In this example, the driver selects the first route (line 502), which corresponds to the highest expected PF regeneration. Based on the driver's selection, it can be inferred that the driver is not under time pressure during this driving cycle and it is possible to schedule the PF regeneration during the driving cycle.When taking a detour where the driver starts on the selected route but does not follow the selected route, as discussed below, the controller may consider the current driver preference to select a route that allows for a higher degree of PF regeneration during destination prediction (based on driving history) and route ranking.

[0064] Back at Fig. 3, the driver's real-time interactions with traffic during the drive cycle may be determined at 320 to determine a current driver state of mind. The driver's state of mind may vary from one state of mind to another based on the driver's interaction with traffic and further based on environmental factors (such as weather) and behavioral factors (such as reaction to a particular situation). The controller may determine the number and frequency of stops made (start-stop frequency) and the duration of each stop. In one example, the driver may be driving along a busy roadway with frequent traffic stops, with the vehicle stopping for several shorter periods of time. In another example, the driver may stop less often but for longer periods of time. As the number of stops increases, the likelihood of completing a regeneration may increase.The driver's frequency and the application force of each of the accelerator and brake pedals may also be determined. In one example, a driver may exhibit a lead foot and accelerate and brake frequently. In another example, a driver may maintain a constant vehicle speed over a period of time without sharp accelerations and decelerations. Additionally, the overall traffic speed (average speed at which other vehicles are traveling on the road) and the spacing between two consecutive vehicles traveling in a lane may be determined. Thus, during lower traffic speeds, when the spacing between consecutive vehicles is smaller, there may be a higher probability that the driver will need to apply the brakes more frequently.In contrast, during higher traffic speeds and when the distance between consecutive vehicles is higher, the vehicle can travel at a constant speed for a longer period of time.

[0065] At 322, the controller may detect possible transitions in the driver's state of mind (e.g., from the initial state S D0 to a state S DK) based on learned driver interactions with traffic and further based on environmental and behavioral factors. In one example, the driver's mood may not change throughout the entire driving cycle, while in another example, the driver's mood may frequently change from one state to another. A non-homogeneous state transition model may be used to determine a probability of transitioning from one mood state to another. In one example, a simplified homogeneous transition matrix (T) may be used to predict possible transitions from one mood state to another.

[0066] An example of a homogeneous transition matrix (T) 700 is shown in Fig. 7. In the homogeneous transition matrix (T), a finite probability is assigned to each transition from one state to another. The probabilities may be based on the driving history (as retrieved from the database) for state transitions while operating on the same route. The finite probabilities in the homogeneous transition matrix may not change based on driver behavior (driver interactions with traffic) or conditions, such as weather, in the current drive cycle.

[0067] As can be seen from the matrix 700, the first row 702 denotes the probabilities of transitioning from a current state of mind of S D0 to the states S D1 and S D2 . If the current state is S D0 is assigned, the probability that the driver will be in state S throughout the entire driving cycle is D0remains, 70%, while the probability that the driver will return to state S at any point during the driving cycle D1 is 20% and the possibility that the driver can switch directly from state S at any point during the driving cycle D0 to state S D2 can transition is 10%. The second row 704 denotes the probabilities of transitioning from a current state of mind of S D1 to the states S D0 and S D2 . If the current state is S D1 is assigned, the probability that the driver will be in state S throughout the entire driving cycle is D1 remains, 50%, while the probability that the driver will return to state S at any point during the driving cycle D0 can go to is 25% and the possibility that the driver can go to state S at any point during the driving cycle D2can transition is 25%. The third row 706 denotes the probabilities of transitioning from a current state of mind of S D2 to the states S D0 and S D1 . If the current state is S D2 is assigned, the probability that the driver will be in state S throughout the entire driving cycle is D2 remains, 50%, while the probability that the driver at any point during the driving cycle goes directly to state S D0 can go to is 25% and the possibility that the driver can go to state S at any point during the driving cycle D1 can be transferred is 25%.

[0068] The transition from one driver's state of mind to another can be estimated based on a dynamic equation. For example, if the driver is in an initial mental state S D0 the transition to a state S Dkbe estimated using equation 2. SDk(t+n)=SDk∗Tn where S Dk is a driver's mood state, t is time, n is the number of possible driver mood states, and T is the homogeneous transition matrix.

[0069] In another example, a non-homogeneous transition matrix (Tn) can be used to predict possible transitions from one state of mind to another. The non-homogeneous matrix can account for real-time driver interactions with traffic and other evolutionary conditions, such as weather, while determining a state transition from an initial driver state to an updated driver state of mind. A non-homogeneous transition model can be experimentally identified and validated using tools including machine learning and exploratory data analysis. Such tools can be used to determine probabilities of transitioning from one state to another. The probabilities of transition for the non-homogeneous transition matrix (Tn) may not be available in precise numbers and may change in real time.In one example, the probability of transitioning from an initial state (p. Dj ) to an updated state (S Dk ), as shown in Equation 3, can be a function of driver interactions with traffic, weather conditions, driver limitations, and time of day. P(S DK |S Dj ) = f (driver interactions with traffic, weather, driving restrictions, time of day) (3)

[0070] Fig. Figure 6 shows a state machine diagram 600 for the transition of a mental state from one state to another. The diagram 600 shows three possible states of mind of the driver, S D0 , S D1 , and S D2 , shown. In an example, if it is determined (based on the driving history) that a driver's state of mind is the first state S D0 the driver can continue in the first state of mind throughout the entire driving cycle. The first state of mind S D0The driver's state of mind is optimal for achieving increased PF regeneration efficiency. Thus, in the first state, the driver can maintain a constant vehicle speed over a period of time without sharp accelerations and decelerations. The driver can continue in the first state if there is no significant change in weather and / or traffic conditions that could cause restrictions and increase the duration of the trip. The probability that the driver remains in the first state of mind is given by P(S D0 |S D0 ) given.

[0071] The state of mind may change from the first state S in response to changes in driver interactions with traffic and / or weather. D0 to the second state S D1 The second state S D2may be a suboptimal state of mind and may have a detrimental effect on PF regeneration efficiency. In one example, changes in driver interactions with traffic may occur due to a change in the weather, such as when it starts to rain heavily while driving. Thus, the driver may accelerate within a short period of time and then frequently apply the brakes to control speed. Frequent application of the brakes may be detrimental to PF regeneration efficiency. In another example, an expected traffic jam may cause the driver to stop longer than planned, and after passing through the traffic jam, the driver may drive aggressively to reach the destination without further delay. The probability of transitioning from the first state to the second state is given by P(S D1 |S D0). Once the driver is in the second state, he or she can continue in the second state for the remainder of the driving cycle. The probability that the driver remains in the second state of mind is given by P(S D1 |S D1 ). After operating in the second state for a period of time, the driver can return to the first state S in response to changes in driver interactions with traffic and / or weather. D0 In one example, due to changes in traffic conditions, such as fewer cars on the road, the driver may begin to drive less aggressively, and there may be a subsequent change in mood.

[0072] The driver's state of mind may change from the second state S in response to changes in driver interactions with traffic and / or weather conditions. D2 to the third state S D2 The third state S D3may be the least optimal state of mind and may adversely affect PF regeneration efficiency. In one example, the driver may frequently stop without significantly accelerating due to further changes in the weather, such as if it starts to snow while driving. In another example, triggers such as a phone call may cause the driver to begin driving aggressively with increased braking frequency. The probability of transitioning from the second state to the third state is given by P(S D2 |S D1 ). Once the driver is in the third state, he or she can continue in the third state for the remainder of the driving cycle. The probability that the driver remains in the third state of mind is given by P(S D3 |S D3). After operating in the third state for a period of time, the driver can switch to the second state S in response to changes in driver interactions with traffic and / or weather. D1 In one example, weather conditions that affect driver interactions may change and the driver may stop less often during the drive cycle. The probability of transitioning from the third state to the second state is given by P(S D1 |S D2 ). The state of mind can also transition from the third state directly to the first state, facilitating improved PF regeneration. In one example, a road blockage may be cleared, allowing the driver to accelerate optimally without frequent braking during the remainder of the drive cycle. The probability of transitioning from the third state to the first state is given by P(S D0 |S D2). Thus, unexpected changes in the traffic situation, such as a road closure, can lead to a transition from the first state to the third state. Due to unexpected changes in traffic, the driver may take an alternative route and driving aggressiveness may also increase. A change in the state from the first to the third state can lead to the termination of an ongoing PF regeneration. The probability of the transition from the first state to the third state is given by P(S D2 |S D1 ). In this way, driver interactions with traffic during the driving cycle and environmental conditions can cause changes in the driver's state of mind.

[0073] Based on the learned driver interactions with traffic during the drive cycle, the controller can determine an updated driver mood. The controller can use the non-homogeneous transition model to determine the updated (current) driver mood. Thus, a variety of changes in the driver mood can occur during the drive cycle (additional moods not represented here), and the controller can continuously update the driver's current mood based on changes in operating conditions, as described above.

[0074] Again at 324, the controller may determine an updated regeneration impact factor corresponding to the updated driver mood. In one example, the controller may use a lookup table to determine the updated regeneration impact factor corresponding to the updated driver mood, where the input is the updated driver mood and the output is the updated regeneration impact factor. A second set of scaling factors may be associated with the updated regeneration impact factor. The total cost function of each route ranked and displayed prior to the change in the driver mood may be reestimated based on changes in the weights due to scaling by the second set of scaling factors.Additionally, in response to the updated driver state of mind, new routes are retrieved from the database and ranked along with the previously displayed routes. Alternatively, the previously displayed routes may be re-ranked and displayed in a different order. Weights scaled by the second set of scaling factors corresponding to the updated regeneration impact factor may be used to estimate a total cost function associated with each of the routes. Estimation of a total cost function based on individual cost functions associated with a route's expected PF regeneration level, the probability of a PF regeneration process to be completed during the drive cycle, the cost of heating the PF to the regeneration temperature, fuel consumption, and the duration of the trip may be performed using Equation 1.The details of the total cost function estimation are given in step 212 of . Fig. 2 described.

[0075] Once the total cost functions for the plurality of available routes between the origin and the destination have been estimated, the routes may be re-ranked based on the total cost function, with the highest-ranked route corresponding to the lowest total cost function. In one example, the ranking of the routes may change after the driver's mood state is updated. In another example, the ranking of the routes may not change after the driver's mood state is updated. At 325, the updated ranking of the routes, as estimated based on the updated total cost functions, may be displayed to the driver.

[0076] Fig. 5B shows a screenshot 550 of an exemplary in-vehicle navigation system, with updated route options displayed following a change in the driver's mood. At the start of the drive cycle, the selected routes between the start and destination were ranked and displayed based on the driver's initial mood, as shown in Fig. 5A. After selecting the first route (from the displayed routes) corresponding to the highest level of achievable PF regeneration, the driver may continue along the selected route. However, after operating on the selected first route for a duration, there may be a change in the driver's mood due to changes in traffic and / or environmental conditions, and based on the driver's updated mood, the previously highest-ranked route may no longer remain the route corresponding to the highest level of achievable PF regeneration. In one example, the change in the driver's mood may be attributed to the driver beginning to drive more aggressively due to the presence of more traffic.Accordingly, the updated routes may include alternative routes with fewer vehicles that are better suited to consistent operation and achieving the desired level of PF regeneration.

[0077] As shown in screenshot 550, for each column: a first box 501 indicates the suggested route between the start and the destination, a second box 503 shows a PF regeneration percentage achievable while driving, and a third box 505 shows the time required to reach the destination from the start. The driver can select one of the four routes via a fourth box included in each row.

[0078] This means that previously displayed routes can be shown again, but the route classification may change. In one example, the route with the second classification can be shown in Fig. 5A is now the highest rated route in Fig. 5B, while the highest rated route is in Fig. 5A as the second classified route in Fig. 5B can be displayed. The route classified as the third in Fig. 5A cannot be displayed as more than one route option. The route classified as the fourth in Fig. 5A can be considered the third classified route in Fig. 5B, while a new route between the current geographic location and the destination may be retrieved from the database and incorporated as a new (fourth ranked) option. In this example, the driver may select the first route corresponding to the highest level of achievable PF regeneration. Since the first route in the updated list differs from the previously selected route, updated navigation instructions may be provided to reroute the driver from the current route. The driver may continue along the newly selected route to the destination. In this way, the ranking of the routes may be updated in real time based on the driver's current state of mind, and the updated ratings may be displayed to the driver.Similar updates to the recommended route list may occur multiple times during the drive cycle as the driver's mood changes based on real-time conditions.

[0079] In this way, an updated driver mood state and a corresponding regeneration factor can be selected from a database by applying a non-homogeneous transition model based on a probability of transitioning from a first driver mood state (and a corresponding first regeneration factor) to an updated driver mood state (and a corresponding updated second regeneration factor), the probability being based on real-time driver interactions with traffic.

[0080] Back at Fig. 3, the routine includes determining at 326 whether a key-off event has occurred, indicating that vehicle operation has stopped. If a key-off event is not indicated, the controller may continue to learn a current driver mood state in real time at 328, and based on the updated driver mood state, route recommendations may be updated in real time. Updating may include updating the ranking of previously displayed routes or displaying new routes (retrieved from the database) to the driver to increase the achievable level of PF regeneration during the given drive cycle.

[0081] At 330, the database may be updated with information learned during the current drive cycle, including driver interactions with traffic, displayed routes, driven road segments, various driver mood states, actual particulate filter regeneration achieved, etc. Furthermore, the circumstances and probabilities of transition (from one state to another), each leading to a change in the driver's mood state, may be incorporated into the update. The details of updating the database with information learned during the current drive cycle are described in Fig. 4 discussed in detail.

[0082] In this manner, a first particulate filter regeneration factor may be selected at the start of a drive cycle based on a vehicle operator's prior driving history, and one or more routes selected from a database may be displayed to the vehicle operator, the one or more routes being ranked based on each of the drive cycle start and destination, corresponding regeneration completion efficiencies, and the first regeneration factor. Navigation instructions for a driver-selected route from the one or more displayed routes may be displayed to the vehicle operator. During the drive cycle, a second particulate filter regeneration factor may be selected based on real-time driver interactions with traffic while driving along the driver-selected route, and route and navigation instructions may be updated accordingly.

[0083] Fig. 4 shows an example method 400 for updating the database of frequently traveled routes. At 402, the routine may include determining whether a vehicle key-on event has occurred. For example, it may be determined whether the vehicle driver has expressed an intent to start vehicle operation. Thus, acknowledging a vehicle key-on event indicates an upcoming vehicle drive cycle. If a vehicle key-on event is not detected, and thus an upcoming vehicle drive cycle is not acknowledged, the method may end and the database may not be updated.

[0084] If the vehicle key-on event is acknowledged, the controller may learn start characteristics, including time and geographic location, of the key-on event at 404. For example, the controller may determine the start characteristics based on information from a vehicle navigation system (e.g., GPS device). In this way, the controller may determine a duration of time the vehicle was stopped at a location (e.g., the starting point) before commencing travel. Furthermore, a duration elapsed since the immediately preceding key-off event may be determined. That is, a stopped duration of the vehicle at the current location may be estimated.

[0085] At 406, the controller may learn details regarding a route of travel of the vehicle, including traveled road segments. This may include topographical information (such as road grade, slope, and terrain) of each of the road segments on the actual route. The details may be learned based on information from the vehicle navigation system and / or from an external server via wireless communication. At 408, the controller may learn details regarding intermediate stops along the way to the destination. The stops may be due to traffic signals, traffic congestion, or the stops may be intentional by the driver. The geographic location and the duration of each of the stops may also be learned. At 410, traffic information for the route, including the number of traffic stops, may be learned via the navigation system.The controller can also determine the speed limits for each road segment and the actual speed of the vehicle's travel.

[0086] At 412, the controller may learn the time of the trip, including a time of day the vehicle is traveling, a date of the trip, what day of the week the vehicle is traveling, etc. At 414, the controller may learn engine operating conditions from a plurality of engine sensors (such as sensor 16 in Fig. 1) learn during each road segment driven, such as engine speed, engine load, engine temperature, etc. In addition, the controller can learn the soot level that has accumulated on the PF while the vehicle is driving.

[0087] At 416, the controller may learn the level of PF regeneration achieved during the current trip. In one example, the level of PF regeneration may be determined based on the change in PF soot level (between the soot level measured at the start of the trip and the end of each road segment), as estimated via an exhaust pressure sensor. In another example, the level of PF regeneration may be determined based on the duration of PF regeneration. It may also learn whether the PF regeneration process terminated prematurely due to adverse operating conditions during the drive cycle. The reasons for the termination of PF regeneration and the engine operating conditions and road conditions at which the PF regeneration event terminated may also be logged.

[0088] At 418, the controller may learn the driver's driving characteristics. These may include, for example, the frequency of brake and accelerator pedal application, the frequency of brake and accelerator pedal release, the transmission shift frequency, the duration of operation in electric mode versus motor mode, etc. The controller may also learn the driver's various mood states and the duration of each mood state during the drive cycle based on the driving characteristics. Furthermore, the conditions (traffic, environment, behavior, etc.) that trigger changes in the driver's mood state may also be learned. Furthermore, for each driver's mood state, the controller may learn the probabilities of transitioning from the current mood state to another depending on the current mood state and the conditions that trigger the change in the driver's mood state.

[0089] At 420, the routine includes determining whether a key-off event has occurred, indicating that vehicle operation has stopped. If a key-off event is not indicated, the controller may continue to collect data related to various aspects of vehicle operation while the vehicle is traveling at 422. If a vehicle stop is confirmed, the method includes learning the target characteristics, including the geographic location of the target, at 424.

[0090] At 425, the controller may learn the fuel consumed during the trip from the start to the destination. As one example, the amount of fuel consumed may be estimated based on the initial and final fuel levels in the fuel tank. In another example, fuel consumption may be estimated based on engine operating conditions. Additionally, the duration of the trip and the time to reach the destination from the start may be learned.

[0091] At 426, the database may be updated with all of the aforementioned data (as collected in steps 404 and 424), including information related to the current route, engine operating conditions, PF regeneration information, and driver driving characteristics. At 428, the current route to travel between the start and destination may be compared to one or more routes (between the start and destination) previously stored in the database. The cost function, as estimated by Equation 1, may be estimated for the current route and compared to the cost function of each of the routes previously stored in the database.At 430, the various routes may be ranked based on the cost function comparison in terms of highest fuel efficiency, shortest travel time, highest level of PF regeneration achieved, and any other cost function selected by the vehicle operator. Furthermore, a Markov chain-based route ranking algorithm may be used. As an example, a route where the target level of PF regeneration may be possible may not be the most fuel-efficient route.

[0092] In this way, the database can be updated with route information, start characteristics, destination characteristics, driver behavior, level of PF regeneration achieved, engine operating conditions, date and time information, and traffic information, etc., upon a vehicle key-off event.

[0093] Fig.Figure 10 shows a prophetic example of the prediction and dynamic selection of a proposed route suitable for optimal particulate filter regeneration.

[0094] A route can be selected from an existing route database based on the engine's PF regeneration requirements, and the current driver state and the selected route can be displayed to the driver. The horizontal (x-axis) represents time, and the vertical markers t1 - t6 indicate key points in the vehicle system's operation.

[0095] At time t1, the driver starts the vehicle (such as a key-on event), and the current geographical location of the vehicle, as determined by an in-vehicle navigation system, is denoted by A. The driver first specifies the geographical location of the destination via an input to the navigation system, as denoted by B. The controller retrieves the driving history (for the driver) from the database, including the driver's characteristics and preferences (such as the frequency of brake use, the average g-force used, the average lane change frequency, etc.). The controller can determine the current day and time of the trip, the starting characteristics (weather conditions and traffic conditions), and based on the retrieved data (the above-mentioned information), the controller assigns the driver an initial state of mind.The controller then selects one or more routes from the database using dynamic programming based on the PF soot level, fuel consumption, travel time, and traffic conditions. Line 1020 shows the change in the soot level deposited on the PF. The dashed line 1021 denotes a threshold above which PF regeneration is desired. Based on the current PF soot level (at time t1), the controller infers that the PF soot level may increase to the threshold soot level during travel from point A to point B and PF regeneration must be performed. The controller then ranks the one or more selected routes as a weighted function of each of a particulate filter regeneration efficiency, a probability of completing a PF regeneration event, fuel efficiency, and travel time.During the classification of one or more selected routes, the weighted function is adjusted based on a regeneration efficiency factor corresponding to the driver's initial state of mind. In this example, the driver's initial state of mind is an optimal state of mind (p. D0), where the particulate filter regeneration efficiency is highest and the probability of completing a PF regeneration event is also increased. Particulate filter regeneration efficiency is an amount of particulate filter regeneration achieved in the drive cycle through a combination of passive and active regeneration. During passive regeneration, soot is burned due to a higher exhaust gas temperature during higher load conditions, and during active regeneration, the temperature of the PF can be increased by flowing electricity therethrough. In response to the impending request for PF regeneration, during the staging of the one or more routes, each is assigned a higher weighting to the particulate filter regeneration efficiency and the probability of completing a PF regeneration event, and a lower weighting is assigned to fuel efficiency and drive time.In this way, the controller displays the route with the highest particulate filter regeneration efficiency and probability of completing a PF regeneration event at the highest rating, and the route with the lowest particulate filter regeneration efficiency and probability of completing a PF regeneration event at the lowest rating. The list is then displayed to the driver, allowing them to select a route based on the rating.

[0096] At time t1, the driver selects route 402, which is the highest-ranked route in the list of suggested routes between start A and destination B, as provided to the driver. However, at time t2, it is observed that the driver deviates from the originally selected route and takes a new route. In response to the change in route, the controller predicts the upcoming route segments between the current location and destination B based on the driver's driving history retrieved from the database and the current traffic conditions. Thus, the routes frequently traveled by the driver during the time of day and / or day of the week are taken into account while predicting the upcoming route segments using stochastic dynamic programming.Based on the driving history and the initial mood, the controller predicts that the driver can take route 404 to a first intermediate stop C and from there route 406 to destination B. The controller then sets the PF regeneration based on the predicted route.

[0097] The driver follows the predicted route and continues along route 404, but it is observed that the driver deviates from route 406 after the first intermediate stop C at time t3. Furthermore, at time t3, based on the driver interactions with traffic between time t1 and t3, the controller updates the driver's mood from the optimal mood to a second, suboptimal mood (S D1). Due to the transition to the second, suboptimal state of mind, it may be learned that the driver can operate the vehicle more aggressively, and driving characteristics, such as increased frequency of braking application, may reduce particulate filter regeneration efficiency and decrease the likelihood of completing a PF regeneration on the given drive cycle. In response to the change in route, the controller once again predicts, based on the driving history and the updated state of mind, that the driver can take route 408 to a second intermediate stop D and from there route 410 to a third intermediate stop E and then route 412 to destination B. However, it is observed that the driver does not take the predicted route segments 408, 410, and 412, but does travel to the destination via a new route 414 and 416.The driver can stop at an intermediate point F on the way to destination B and ultimately reach the destination at time t4.

[0098] It is observed that the soot level in the PF reaches the threshold level 421 at time t3, and passive regeneration of the soot is performed. However, because the driver does not follow a suggested route optimal for PF regeneration, and due to the driver's suboptimal mood, the level of PF regeneration achieved upon reaching destination B at time t4 is lower than the level of PF regeneration that could have been achieved if the driver had taken route 1002 from start A to destination B. Since the driver deviates from the suggested and predicted routes, setting passive regeneration of the PF (between time t3 and t4) may also not be possible.The dotted line 1022 shows a possible change in the soot level in the PF that would have occurred if route 1002 had been taken, assuming the driver's mood had not changed from the first to the second. As observed from traces 1020 and 1022, the amount of soot burned between time t3 and t4 would be higher relative to the amount of soot burned during the journey via routes 1014 and 1016 if the proposed route 1002 had been taken.

[0099] At time t4, upon reaching destination B (at the vehicle key-off event), the database can be updated with route information, starting characteristics (such as geographical location), destination characteristics, the location of each stop taken, driver interactions with traffic (such as gear shift frequency, pedal application and release frequency, brake application frequency, etc.), level of PF regeneration achieved, engine operating conditions (such as engine speed, engine load, engine temperature, etc.), date and time information, driver mood, and traffic information. In addition, information on road grade, terrain, and slope for each segment of the route can be integrated into the database. The data stored in the database can be used for future route selection and / or prediction.

[0100] In this way, PF regeneration can be effectively scheduled and performed by dynamically selecting a route based on a particulate filter regeneration request during a drive cycle from a variety of routes available in a database. By considering a current driver mood, representative of the driver's real-time driving behavior, the likelihood of a driver selecting a recommended route with higher PF regeneration efficiency is increased. By estimating the driver's mood in real time based on driver interactions with traffic and environmental conditions, the ranking of navigation routes corresponding to the probability of achieving a desired level of PF regeneration can be updated as driver behavior changes.By predicting a destination or segments of an upcoming route based on driving history and driving statistics stored in the route database, it may be possible to schedule PF regenerations even during trips where a final destination has not been specified by the driver or when the driver deviates from a selected route.The technical effect of maintaining a database of driver mental states and frequently driven routes with information, including the potential level of PF regeneration achievable on each route, is that an initial mental state at a vehicle key-on event can be selected from the database based on the driving history, and a route can be selected from the database based on the PF soot level and the driver-selected cost function, including the highest fuel efficiency and the lowest travel time for the drive cycle. In this way, PF soot overload during higher than threshold PF soot loading can be reduced by selecting a favorable route for opportunistic PF regeneration, thereby improving engine performance.By allowing a PF to be regenerated opportunistically using passive regeneration, the need for active regeneration is reduced, thereby providing additional fuel efficiency benefits.

[0101] An exemplary engine method includes learning, after each drive cycle, a particulate filter regeneration efficiency dependent on one or more characteristics of a traveled route and a vehicle operator's behavior along the route; updating a database based on the learning; and, at the start of a drive cycle, displaying one or more routes selected from the database to a vehicle operator, the selection based on a particulate filter soot load at the start of the drive cycle. In a preceding example, the selection is further additionally or optionally based on a vehicle operator-specified target for the drive cycle relative to a start of the drive cycle.In any or all of the preceding examples, the selection is further additionally or optionally based on a cost function selected by the vehicle operator, including one or more of a highest fuel efficiency and a least travel time for the drive cycle. Any or all of the preceding examples further additionally or optionally include ranking the one or more routes as a weighted function of each of a particulate filter regeneration efficiency, a probability of completing a particulate filter regeneration event, a fuel efficiency, and a travel time of each of the one or more routes.In any or all of the preceding examples, the particulate filter regeneration efficiency additionally or optionally includes a degree of particulate filter regeneration predicted for the drive cycle, and wherein the one or more characteristics of the traveled route include a number of stops, road gradient, and traffic conditions. In any or all of the preceding examples, each of the particulate filter regeneration efficiency and the probability of completing a particulate filter regeneration event is additionally or optionally assigned higher weights when the particulate filter soot load is higher than a threshold. In any or all of the preceding examples, each of the fuel efficiency and the drive time is additionally or optionally assigned higher weights when the particulate filter soot load is lower than the threshold.Any or all of the preceding examples additionally or optionally include, in response to the vehicle operator not selecting a route from the one or more routes displayed to the vehicle operator, dynamically predicting an upcoming route segment based on a driving history of the vehicle operator retrieved from the database. In any or all of the preceding examples, the driving history of the vehicle operator additionally or optionally includes routes previously driven depending on one or more of the time of day, day of week, and traffic conditions. Any or all of the preceding examples additionally or optionally include, in response to the vehicle operator not selecting a route from the one or more routes displayed to the vehicle operator, dynamically predicting an upcoming route segment based on a driving history of the vehicle operator retrieved from the database.additionally or optionally include, in response to the vehicle operator selecting a route from the one or more routes displayed to the vehicle operator, initiating travel along the route and then departing from the selected route, wherein an upcoming route segment is dynamically predicted based on a driving history of the vehicle operator retrieved from the database. Any or all of the preceding examples further additionally or optionally include, based on the route selected by the vehicle operator from the one or more displayed routes, determining an active particulate filter regeneration event during the drive cycle, wherein a temperature of the particulate filter is increased during the active particulate filter regeneration event by flowing electrical current through the particulate filter. Any or all of the preceding examples further additionally or optionally include,further additionally or optionally include, if a destination for the drive cycle is not specified by the vehicle operator, predicting the destination based on a driving history of the vehicle operator as retrieved from the database, and displaying to the vehicle operator one or more routes selected from the database based on the predicted destination. In any or all of the preceding examples, the start of the drive cycle additionally or optionally includes a vehicle key-on event, and wherein updating the database based on the learning at a vehicle key-off event includes updating the database with route information, start characteristics, destination characteristics, vehicle operator behavior, level of particulate filter regeneration achieved, engine operating conditions, date and time information, and traffic information.

[0102] Another example engine method includes: in response to a vehicle operator destination selection indicated via a display of a vehicle, estimating an exhaust particulate filter soot load; determining a current location of the vehicle; retrieving one or more routes from the current location to the destination from a database; ranking the one or more routes based on a particulate filter regeneration efficiency, fuel efficiency, and travel time of each route; and displaying to the vehicle operator the one or more routes to the selected destination in order of ranking. Any of the preceding examples further additionally or optionally includes, in response to a greater than threshold particulate filter soot load, ranking the one or more routes by assigning a higher weight to the particulate filter regeneration efficiency and a lower weight to each of the fuel efficiency and travel time.Any or all of the preceding examples further additionally or optionally include, in response to less than a threshold particulate filter soot load, ranking the one or more routes by assigning the higher weighting to each of fuel efficiency and travel time and assigning the lower weighting to particulate filter regeneration efficiency. Any or all of the preceding examples further include,further additionally or optionally include, in response to the vehicle operator not selecting a route from the one or more displayed routes or the vehicle operator deviating from a selected route, predicting one or more route segments from the current location to the destination based on the vehicle operator's driving history retrieved from the database, and determining passive regeneration of the particulate filter based on the one or more predicted route segments.

[0103] In yet another example, a vehicle method comprises: in response to no destination for a drive cycle being specified by a vehicle operator, determining a current location of the vehicle, retrieving a driving history of the vehicle operator from a database, predicting a destination based on the driving history, dynamically updating the selection of one or more upcoming route segments based on the current location of the vehicle relative to the predicted destination, ranking the one or more upcoming route segments based on each of a corresponding particulate filter regeneration efficiency, fuel efficiency, and drive time, and displaying to the vehicle operator the one or more upcoming route segments to the predicted destination hierarchically in order of ranking.Any upcoming example further additionally or optionally includes, in response to the vehicle operator not selecting a route from the one or more displayed upcoming route segments, dynamically updating the one or more upcoming route segments based on driving history and determining active particulate filter regeneration based on the one or more updated segments. Any or all of the preceding examples further additionally or optionally include, upon completion of a drive cycle, updating the database with route segment information, including vehicle operator behavior, level of achieved particulate filter regeneration, engine operating conditions over the completed drive cycle, date and time of trip information, and traffic information over the drive cycle.

[0104] In another example, an engine method comprises: at a start of a drive cycle, displaying a first driving route in response to each of a particulate filter (PF) loading and a previous driving history; and while driving along the first driving route, displaying an updated route in response to each of traffic conditions and a comparison of the real-time driving history along the first route on the drive cycle relative to the previous driving history. In any preceding example, displaying the first driving route additionally or optionally includes selecting the first driving route from a database including a plurality of driving routes based on a first inferred driver mood, wherein the inferred driver mood is based on the previous driving history.and wherein displaying the updated route includes selecting the updated route from the database based on an updated driver mood state. In any or all of the preceding examples, the updated driver mood state is additionally or optionally selected from a plurality of derived driver mood states stored in the database, wherein the updated driver mood state is selected based on a comparison of the real-time driving history along the first route on the drive cycle relative to the prior driving history, wherein each of the plurality of derived driver mood states has an associated PF regeneration factor. In any or all of the preceding examples, displaying the first route includes, in response to a vehicle operator-specified target for the drive cycle,additionally or optionally further displaying one or more routes retrieved from the plurality of driving routes included in the database, the one or more routes ranked as a first function of the corresponding PF regeneration efficiency, a probability of completing a PF regeneration event during the drive cycle, and a first PF regeneration factor associated with the first driver state of mind, wherein the corresponding PF regeneration efficiency for each of the one or more routes is determined depending on the corresponding degree of PF regeneration predicted for the drive cycle and the first PF regeneration factor. In any or all of the preceding examples, displaying the updated route, in response to the comparison, additionally or optionally includes displaying the one or more routes, the one or more routes,which are ranked as a second function of the updated PF regeneration efficiency, the probability of completing the PF regeneration event during the drive cycle, and a second PF regeneration factor associated with the updated driver state of mind, wherein the corresponding PF regeneration efficiency for each of the one or more routes is determined as a function of the corresponding degree of PF regeneration predicted for the drive cycle and the second PF regeneration factor. In any or all of the preceding examples, ranking the one or more routes as the function of the first PF regeneration factor includes ranking the one or more routes based on assigned weights for each of the PF regeneration efficiency and the probability of completing the PF regeneration event during the drive cycle.wherein the assigned weights are ranked by a first function based on the first PF regeneration factor, and wherein ranking the one or more routes as the function of the updated PF regeneration factor includes ranking based on assigned weights in each of the PF regeneration efficiency and the probability of completing the PF regeneration event during the drive cycle, wherein the assigned weights are ranked by a second set of factors corresponding to the second PF regeneration factor, the second set being different from the first set. In any or all of the preceding examples, the real-time driving history additionally or optionally includes real-time driver interactions with traffic, including real-time accelerator usage and real-time brake usage during the drive cycle while driving along the first driving route.and wherein the previous driving history includes a frequency of brake usage, the average acceleration force used, and the average lane change frequency while driving along the first route in one or more drive cycles prior to the drive cycle. Any or all of the preceding examples further additionally or optionally include learning, upon completion of the drive cycle, a level of PF regeneration achieved during the drive cycle and then updating the database with the achieved learned level of PF regeneration for the drive cycle, the first derived driver state of mind, and the updated driver state of mind.

[0105] In yet another example, a method comprises: at a start of a drive cycle, selecting a first particulate filter regeneration factor based on a vehicle operator's prior driving history; displaying, to the vehicle operator, one or more routes selected from a database, the one or more routes ranked based on each of the drive cycle start and destination, the corresponding regeneration completion efficiencies, and the first regeneration factor; displaying, to the vehicle operator, navigation instructions for a route selected by the vehicle operator from the one or more displayed routes; and during the drive cycle, selecting a second particulate filter regeneration factor based on real-time driver interactions with traffic while driving along the vehicle operator-selected route.In any preceding example, each of the first regeneration factor and the second regeneration factor is additionally or optionally selected from a plurality of regeneration factors stored in the database, wherein each of the plurality of regeneration factors corresponds to a different driver state of mind. In any or all of the preceding examples, selecting the second regeneration factor additionally or optionally includes applying a non-homogeneous transition model to select the second regeneration factor from the plurality of regeneration factors based on a probability of transitioning from the first regeneration factor to the second regeneration factor, wherein the probability is based on real-time driver interactions with traffic.In any or all of the preceding examples, the driver's past driving history additionally or optionally includes routes driven by the vehicle operator depending on one or more of a time of day, day of week, the start and finish of the drive cycle, and driving characteristics, including brake usage frequency, average g-force used, and average lane change frequency. In any or all of the preceding examples, the real-time driver interactions with traffic additionally or optionally include one or more of stop frequency, lane change frequency, accelerator pedal input, and brake input during the drive cycle.In any or all of the preceding examples, selecting the first regeneration factor is further additionally or optionally based on traffic conditions at the start of the drive cycle and environmental conditions at the start of the drive cycle, including ambient temperature, ambient humidity, and precipitation. In any or all of the preceding examples, the one or more routes being ranked further additionally or optionally include ranking each of the one or more routes based on a weighted function of each of the corresponding regeneration completion efficiencies, a probability of completion of a particulate filter regeneration event, a fuel efficiency, and a time to destination of each of the one or more routes, wherein the weighted function is ranked based on the first regeneration factor. Any or all of the preceding examples includefurther additionally or optionally comprising, in response to selecting the second particulate filter regeneration factor, updating the weighted function that is ranked based on the second regeneration factor and then updating the ranking of the one or more routes.

[0106] In another further example, a vehicle system comprises: a vehicle, a navigation system wirelessly connected to an external network, a display, an engine including an intake system and an exhaust system, the exhaust system including a particulate filter (PF) coupled to an exhaust passage and a pressure sensor coupled to the exhaust passage upstream of the particulate filter, and a controller having computer-readable instructions stored in non-transitory memory for: at a start of a drive cycle, displaying a first route based on PF loading and a first driver mood, and in response to driver interactions with traffic while traveling on the first route, displaying a plurality of updated routes based on a second driver mood,wherein the first driver mood state is selected from a database based on each of the PF load and a driving history, and a change from the first driver mood state to the second driver mood state is based on the driver interactions with traffic while driving on the first route. In any preceding example, the first route is additionally or optionally selected based on a first weighted PF regeneration efficiency, wherein the first weighted PF regeneration efficiency is based on a first PF regeneration factor corresponding to the first driver mood state. In any or all of the preceding examples, the plurality of updated routes is additionally or optionally selected based on a second weighted PF regeneration efficiency, wherein the second weighted PF regeneration efficiency is based on a second PF regeneration factor.corresponding to the second driver state of mind, and wherein displaying the plurality of updated routes includes ranking each of the plurality of updated routes based on the second weighted PF regeneration efficiency. In any or all of the preceding examples, the controller additionally or optionally includes further instructions for: learning, during the drive cycle, driver interactions with traffic, displayed routes, driven road segments, the driver state of mind, the achieved particulate filter regeneration, and, upon completion of the drive cycle, updating the database based on the learning.

[0107] In another representation, the vehicle is a hybrid vehicle system. In a preceding example, an example method for a hybrid vehicle additionally or optionally includes: at a start of a drive cycle, selecting a first value indicative of a driver's state of mind based on a driving history; displaying, to the driver, one or more routes selected from a database, wherein a regeneration factor of each of the one or more routes is based on the first value; and during the drive cycle, selecting, in real time, a second value indicative of an updated driver's state of mind based on real-time driver interactions with traffic.

[0108] It should be noted that the control and estimation routines contained herein may be used with various engine and / or vehicle system configurations. The control methods and routines disclosed herein may be stored as executable instructions in non-volatile memory and may be executed by the control system, including the controller in combination with the various sensors, actuators, and other engine hardware. The specific routines described herein may represent one or more of any number of processing strategies, such as event-driven, interrupt-driven, multitasking, multithreading, and the like. Accordingly, various illustrated acts, operations, and / or functions may be performed in the illustrated sequence or in parallel, or in some cases, may be omitted.Likewise, the processing order is not necessarily required to achieve the features and advantages of the exemplary embodiments described herein, but rather is provided for ease of illustration and description. One or more of the illustrated acts, operations, and / or functions may be performed repeatedly depending on the particular strategy employed. Further, the described acts, operations, and / or functions may graphically represent code to be programmed into non-transitory memory of the computer-readable storage medium in the engine control system, wherein the described acts are performed by executing the instructions in a system including the various engine hardware components in combination with the electronic controller.

[0109] It is understood that the configurations and routines disclosed herein are exemplary in nature, and these specific embodiments are not to be considered limiting, as numerous variations are possible. For example, the above technique may be applied to V-6, I-4, I-6, V-12, horizontally opposed 4-cylinder, and other engine types. The subject matter of the present disclosure includes all novel and non-obvious combinations and subcombinations of the various systems and configurations, and other features, functions, and / or characteristics disclosed herein.

[0110] The following claims particularly set forth certain combinations and subcombinations that are considered novel and non-obvious. These claims may refer to "a" element or "a first" element, or the equivalent thereof. Such claims should be understood to encompass the inclusion of one or more such elements and neither require nor exclude two or more such elements. Other combinations and subcombinations of the disclosed features, functions, elements, and / or properties may be claimed by amending the present claims or by filing new claims in this or a related application.Such claims, whether broader, narrower, equal, or different in scope than the original claims, are also considered to be included within the subject matter of the present disclosure.

Claims

[1] Method comprising: at the start of a driving cycle, displaying, by an electronic control, a first driving distance in response to each of a particulate filter (PF) loading and a previous driving history; and while driving along the first route, displaying, by the electronic control, an updated route in response to each of traffic conditions and a comparison of a real-time driving history along the first route on the drive cycle relative to the previous driving history. [2] The method of claim 1, wherein displaying the first route includes selecting the first route from a database including a plurality of routes based on the first inferred mood state of the driver, wherein the inferred mood state of the driver is based on the previous driving history, and wherein displaying the updated route includes selecting the updated route from the database based on an updated mood state of the driver. [3] The method of claim 2, wherein the updated driver mood state is selected from a plurality of derived driver mood states stored in the database, the updated driver mood state being selected based on the comparison of the real-time driving history along the first route on the drive cycle relative to the previous driving history, each of the plurality of derived driver mood states having an associated PF regeneration factor. [4] The method of claim 3, wherein displaying the first travel route, in response to a vehicle operator specified target for the drive cycle, includes displaying one or more routes retrieved from the plurality of travel routes included in the database, the one or more routes ranked as a first function of the corresponding PF regeneration efficiency, a probability of completing a PF regeneration event during the drive cycle, and a first PF regeneration factor associated with the first derived driver state of mind, wherein the corresponding PF regeneration efficiency for each of the one or more routes is determined as a function of a corresponding level of PF regeneration predicted for the drive cycle and the first PF regeneration factor. [5] The method of claim 4, wherein displaying the updated distance, in response to the comparison, includes displaying the one or more distances, the one or more distances ranked as a second function of the updated PF regeneration efficiency, the probability of completion of the PF regeneration event during the drive cycle, and a second PF regeneration factor associated with the updated driver state of mind, wherein the corresponding PF regeneration efficiency for each of the one or more distances is determined as a function of the corresponding level of PF regeneration predicted for the drive cycle and the second PF regeneration factor. [6] The method of claim 5, wherein ranking the one or more routes as a function of the first PF regeneration factor includes ranking the one or more routes based on assigned weights for each of the PF regeneration efficiency and the probability of completing the PF regeneration event during the drive cycle, the assigned weights being ranked by a first function based on the first PF regeneration factor, and wherein ranking the one or more routes as a function of the updated PF regeneration factor includes ranking based on assigned weights for each of the PF regeneration efficiency and the probability of completing the PF regeneration event during the drive cycle, the assigned weights being ranked by a second set of factors corresponding to the second PF regeneration factor,where the second sentence differs from a first sentence., [7] The method of claim 1, wherein the real-time driving history includes real-time driver interactions with traffic, including real-time accelerator pedal usage and real-time brake usage during the drive cycle while driving along the first travel route, and wherein the past driving history includes a frequency of brake usage, average g-force used, and average lane change frequency while driving along the first travel route in one or more drive cycles prior to the drive cycle. [8] The method of claim 2, further comprising learning, at the completion of the drive cycle, a level of PF regeneration achieved during the drive cycle, and then updating the database with the achieved learned level of PF regeneration for the drive cycle, the first inferred driver mood, and the updated driver mood. [9] Vehicle system comprising: a vehicle; a navigation system that is wirelessly connected to an external network; an advertisement; an engine including an intake system and an exhaust system, the exhaust system including a particulate filter (PF) coupled to an exhaust passage and a pressure sensor coupled to the exhaust passage upstream of the particulate filter; and a controller having computer-readable instructions stored in non-volatile memory for the following: at the start of a drive cycle, selecting a first PF regeneration factor based on a previous driving history of a vehicle driver; Displaying, to the vehicle operator, one or more driving routes selected from a database, the one or more routes being ranked based on each of the drive cycle start and finish, the corresponding regeneration completion efficiencies, and the first PF regeneration factor; Displaying, to the driver, navigation instructions for a route selected by the driver from the one or more displayed routes; and during the drive cycle, selecting a second PF regeneration factor based on real-time driver interactions with traffic while driving along the driver-selected route. [10] The system of claim 9, wherein each of the first PF regeneration factor and the second PF regeneration factor is selected from a plurality of PF regeneration factors stored in the database, each of the plurality of PF regeneration factors corresponding to a different state of mind of the driver. [11] The system of claim 10, wherein selecting the second PF regeneration factor includes applying a non-homogeneous transition model to select the second PF regeneration factor from the plurality of PF regeneration factors based on a probability of transitioning from the first PF regeneration factor to the second PF regeneration factor, the probability being based on real-time driver interactions with traffic. [12] The system of claim 9, wherein the driver's past driving history includes routes driven by the vehicle operator as a function of one or more of a time of day, day of week, drive cycle start and destination, and driving characteristics including frequency of brake use, average g-force used, and an average lane change frequency, wherein the real-time driver interactions with traffic include one or more of frequency of stops, frequency of lane changes, accelerator input, and brake input during the drive cycle. [13] The system of claim 9, wherein selecting the first PF regeneration factor is further based on traffic conditions at the start of the drive cycle and environmental conditions at the start of the drive cycle, including ambient temperature, ambient humidity, and precipitation. [14] The system of claim 9, wherein the one or more routes being ranked further include ranking each of the one or more routes based on a weighted function of each of the corresponding regeneration completion efficiencies, a probability of completing a particulate filter regeneration event, a fuel efficiency, and a time to destination of each of the one or more routes, the weighted function being ranked based on the first PF regeneration factor. [15] The system of claim 14, wherein the controller further includes instructions for: in response to selecting the second PF regeneration factor, updating the weighted function that is ranked based on the second PF regeneration factor, and then updating the ranking of the one or more routes.

Images