Patent Application
20130274952
The present disclosure relates to vehicle powertrain control and, more specifically, to predictive vehicle powertrain control based on driving history.
Motorized vehicles include a powertrain operable to propel the vehicle and power the onboard vehicle electronics. The powertrain typically includes an engine that powers the final drive system through a multi-speed transmission. Many of today's conventional, gas-powered vehicles are powered by an internal combustion (IC) engine.
Hybrid vehicles have been developed and continue to be developed. Conventional hybrid electric vehicles (HEVs) combine internal combustion engines with electric propulsion systems to achieve better fuel economy than non-hybrid vehicles. Plugin hybrid electric vehicles (PHEVs) share the characteristics of both conventional hybrid electric vehicles and all-electric vehicles by using rechargeable batteries that can be restored to full charge by connecting (e.g. via a plug) to an external electric power source.
Despite the introduction of hybrid vehicles and improved conventional gas powered vehicles, the automotive industry is continually faced with the challenge of improving fuel economy and reducing emissions without sacrificing vehicle performance. As mentioned above, there are many different types of vehicles in existence today with numerous others being developed for the future. Accordingly, there is a need and desire for a technique for improving fuel economy and reducing emissions without sacrificing vehicle performance that will work with many different types of vehicles.
In one form, the present disclosure provides a method of controlling a vehicle powertrain. The method comprises determining a present location of the vehicle; predicting a route of travel for the vehicle from the present location based on the current day and time; predicting powertrain loads and speeds based on the predicted route of travel; and optimizing a powertrain operation based on the predicted powertrain loads and speeds.
The present disclosure also provides a powertrain apparatus for a vehicle. The apparatus comprises a controller adapted to determine a present location of the vehicle; predict a route of travel for the vehicle from the present location based on the current day and time; predict powertrain loads and speeds based on the predicted route of travel; and optimize a powertrain operation based on the predicted powertrain loads and speeds.
In one embodiment, the optimized powertrain operation comprises one of shift scheduling and battery control.
In another embodiment, predicting the route of travel comprises determining if a next segment in a map database associated with a segment corresponding to the present location is traveled more than a predetermined threshold on a similar day and time as the current day and time; and adding the next segment to the predicted route of travel if it is determined that the next segment is traveled more than the predetermined threshold on a similar day and time as the current day and time. In another embodiment, the next segment is not added if additional information indicates that another segment should be added to the predicted route. Additional information may be input from a vehicle to vehicle data source, a vehicle to infrastructure data source, or a navigation system.
Further areas of applicability of the present disclosure will become apparent from the detailed description and claims provided hereinafter. It should be understood that the detailed description, including disclosed embodiments and drawings, are merely exemplary in nature intended for purposes of illustration only and are not intended to limit the scope of the invention, its application or use. Thus, variations that do not depart from the gist of the invention are intended to be within the scope of the invention.
According to the principles disclosed herein, and as discussed below, predictive control of the powertrain of various conventional and hybrid vehicles can be performed to improve fuel economy and emissions using predicted vehicle usage based on the vehicle's driving history. According to the principles disclosed herein, driving history, GPS location and map information are used to predict future loads and speeds for the current trip. In addition, driver inputs will not be required for the method and system disclosed herein to make a prediction of the driver's route and intended destination.
Predicting the loads and speeds of the vehicle for the duration of a trip allows shift scheduling to be performed on conventional vehicles and allows transmission and battery control to be performed on HEVs and PHEVs. For example, predicted trip load and grades can be used to optimize battery charging and discharging locations along the trip. Moreover, the modes of transmission operation (e.g., in electrically variable transmissions) or gear ratio selection on conventional and other HEVs can be optimized. For PHEVs, which are generally designed to operate in two modes (a charge depleting mode or a charge sustaining mode), prediction of battery charging locations can be used to change battery discharging strategy (in the charge depleting mode) to be more or less aggressive.
As will be shown below, GPS and map data for road segments traveled are stored with a time stamp in a non-volatile memory. As used herein, a “road segment” is derived from how navigation systems truncate any road into specific little blocks or segments. Generally, nothing changes on a segment (i.e., a segment will contain a constant speed and direction, there will be no intersections, stop signs, etc.). During vehicle operation, a prediction concerning the next probable on-coming road segment is made based on the relative weighting of several factors such as e.g., history versus function class (i.e., any information from sources other than the previously stored road segment or history), type of road, straightness of the segment, etc. For example, if travel history is weighted heavily, and at the end of segment X there is a history of immediate travel on segment Y, then the probability of travel on segment Y on a route having segment X is large; thus, segment Y will be considered part of the route being traveled. However, if the history information indicates that segments A, B, C, D and E are travelled next, but information from another source (e.g., a vehicle to vehicle source) indicates that segment C is not a good segment to travel and proposes an alternative segment K, segment K will be added to the route if function class is weighed heavily (travel is now changed to segments A, B, K, D, and E instead off segments A, B, C, D and E); if history is weighed heavily, however, the projected route will remain as segments A, B, C, D and E.
As another example, if the vehicle has a history of stopping after segment Z, and the time stamp for segment Z matches closely with the current time, day of week, etc., then the probability of travel after segment Z is low. Once the probability of further travel falls below a calibrated threshold, the route is assumed to be completed and the destination is presumed to have been reached. If the trip continues past segment Z, however, the segment predictions will begin again.
As will be discussed below with reference to
As can be seen in
If it is determined that the next segment from the map database was traveled with a high frequency (i.e., greater than the predetermined percentage) on a similar past day and time, the method 100 includes the next segment in the current route (step 108) and continues at step 106 to check another “next” segment. Thus, steps 106 and 108 add segments to the predicted route based on route segments (and time stamps) previously stored in the map database.
If at step 106 it was determined that the next segment from the map database was not traveled with a high frequency on a similar past day and time, the method continues at step 110 where stored route information (Le., previously traveled segments) corresponding to the predicted route is retrieved from the map database. The route information preferably includes historical battery charging locations, which are also factors for optimizing the powertrain of PHEVs and similar vehicles. As mentioned above, the prediction of battery charging locations can be used to change battery discharging strategy to be more or less aggressive. At step 112, the method 100 predicts powertrain loads and speeds using a powertrain and vehicle model and the retrieved route information. Using the predicted powertrain loads and speeds, at step 114, an optimized powertrain control strategy is then developed for the type of vehicle. For conventional vehicles, this means that e.g., shift maps for a shifting schedule can be modified. For HEVs and PHEVs, battery charging and discharging scheduling can be modified based on the desired aggressiveness of the schedule. Relevant engine commands needed to implement the new strategy are also developed. The predictive powertrain control strategy is executed at step 116.
It should be appreciated that the disclosed system 10 and method 100 enhance the real world fuel economy of the vehicle, allowing the vehicle's owner to save money on fuel. Better fuel economy is also beneficial to the environment because less fuel is being consumed and less emissions are entering the atmosphere. The disclosed system 10 and method 100 capitalize on information that is readily available from onboard components and systems already present within the vehicle. As such, the system 10 and method 100 are easily and inexpensively implemented into the vehicle. Moreover, the system 10 and method 100 disclosed herein do not require the driver to enter a route or other information to successfully operate and improve the vehicle's fuel economy.
This application claims the benefit of U.S. Provisional Serial No. 61/624,512, filed Apr. 16, 2012.
This invention was made, at least in part, under U.S. Government, Department of Energy, Contract No. DE-EE0002720. The Government may have rights in this invention.
| Number | Date | Country | |
|---|---|---|---|
| 61624512 | Apr 2012 | US |
