Combined quasi-static backward modeling and look-ahead fuzzy control of vehicles

https://doi.org/10.1016/j.eswa.2011.07.012Get rights and content

Abstract

Vehicle modeling can play an important role in vehicle power train design, control and energy management investigation. This paper presents a method for vehicle power train modeling. The key feature of the method is its presentation of the dynamic of vehicle based on the road information. This ability makes the method suitable for look-ahead energy management and fuel economy optimal control problems. With the aid of a road slope database, road geometry ahead of the vehicle is extracted. A fuzzy controller is developed that receives this information and controls the velocity of the vehicle with respect to its fuel consumption. In order to maintain the operation of the combustion engine near its efficient region, the fuzzy controller commands a continuously variable transmission. Simulations are carried out using real road data. The results are presented and discussed.

Highlights

► A drive train model which considers look-ahead trajectory characteristics is presented. ► A fuzzy controller which manages velocity with respect to fuel consumption is developed. ► Control of continuously variable transmission helps combustion engine operate in its efficient region.

Introduction

Improving fuel efficiency in motor vehicles can help reduce the environmental impacts associated with the vehicles emissions. Several techniques can be used to improve fuel efficiency in vehicles including:

  • (i)

    Down weighting: Without compromising vehicle safety and performance, some steel parts of vehicle can be replaced with light and high strength materials.

  • (ii)

    Reducing rolling resistance: Using efficient tyres, rolling resistance can be reduced.

  • (iii)

    Reducing aerodynamic resistance: Employing vehicle shapes with less aerodynamic drag can reduce the wind resistance and, thus fuel consumption.

  • (iv)

    Downsizing engine: Using smaller engines, which tend to work closer to optimal operation, would reduce fuel consumption. Recent improvement in engine technology allows engine downsizing without losing the power.

  • (v)

    Choosing advanced transmission technology: Utilizing new transmission technologies improves the efficiency and consequently diminishes fuel consumption.

  • (vi)

    Increasing electrification: Using advanced battery technology leads to vehicles that can run accessories without dragging power from crankshaft and, consequently the battery can be recharged in lower load conditions helping engine operate efficiently.

  • (vii)

    Advanced drive train: Using two sources of energy in hybrid electric vehicles and storing the brake energy causes significant improvement on fuel economy.

Over the past few decades, there has been significant development in automobile engine and body technologies. Therefore, it would not be easy to achieve key gains in fuel economy through modifications of the vehicle engine and body. However, the advent of modern control strategies and the growing environmental concerns associated with vehicles provide new opportunities to enhance fuel economy in vehicles. For example, an advanced automatic transmission can offer a greater control of the engine. Additionally, the combination of road information and fuzzy control system can enable the engine to operate more efficiently over a wider variety of speeds. To improve fuel economy, a method is presented which incorporates the stated advantages in a model that is developed for a conventional vehicle.

The simulation and computer modeling approach can be employed to decrease the expense of a product from design to prototyping and mass production. Improving fuel economy, performance and drivability in vehicle design has resulted in an increase in the number of the simulation tools either in commercial divisions or in academic communities. Several simulation tools such as simple electric vehicle simulation (SIMPLEV) (Cole, 2004) from the DOE’s Idaho National Laboratory, MARVEL and PSAT (http://www.transportation.anl.gov/modeling_simulation/PSAT/) from Argonne National Laboratory, CarSim (http://www.carsim.com) from AeroVironment Inc., ADVISOR (Wipke, Cuddy, & Burch, 1999) from the DOE’s National Renewable Energy Laboratory, Vehicle Mission Simulator (Noons, Swann, & Green, 1998), and others (Gao, Mi, & Emadi, 2007) have been developed that can model the operation of conventional and hybrid electric vehicle power trains. In all these software tools, the driver demand is determined by “drive cycle”. Drive cycle specifies the vehicle speed in a predefined pattern. The road slope is defined by a constant grade in the model. Among the software tools, ADVISOR and PSAT are more popular than the others. ADVISOR is a backward facing program in which the direction of calculation of tractive effort starts from wheels and move toward engine (Markel et al., 2002, Wipke et al., 1999). Backward model is faster in term of simulation time. PSAT, on the other hand, is a forward looking simulator which models a vehicle in the same way as a real vehicle works, and the calculation direction starts from the driver demand. The forward looking modeling can to be used as a hardware-in-the-loop/software-in-the-loop testing program (http://www.transportation.anl.gov/modeling_simulation/PSAT/). One of the issues with the existing software tools is that their modification to incorporate different control strategies such as optimal control methods is not simple. Furthermore, when the real circumstances and different loads such as wind and road slope changes come around, these programs become ineffective.

The driving behavior, driving pattern, and road topology affect the fuel consumption of a vehicle. From the standpoint of energy management and optimal flow of energy, the knowledge about disturbances relating to driving route, traffic and road geometry can help in the development of a suitable strategy. Such a strategy is known as look-ahead control or look-ahead energy management. The look-ahead control is a “predictive control scheme with additional knowledge about some of the future disturbances on the road topography ahead of the vehicle” (Hellström, Ivarsson, Ãslund, & Nielsen, 2009). Developing a control algorithm, using look-ahead information, allows planning how and when to speed up and slow down the vehicle. Dynamic programming (DP) and an on-board load database were used by Hellström et al. (2009) to control the speed of a truck before uphills or downhills. The simulations showed that a considerable reduction of the fuel consumption is achievable, i.e. 3.5% fuel economy on the 120 km route respect to the trip time. However, DP is time intensive, so its application to real-time systems is limited. Global positioning techniques were applied to determine the road geometric characteristics. Han and Rizos (1999) used GPS data and Kalman filtering to determine the road slope and height to support a solar car race strategy.

The concept of look–ahead control has been expressed as adaptive or intelligent cruise control in some studies (Abdullah et al., 2008, García-Ortiz et al., 1995, Ioannou et al., 1993, Kamble et al., 2009). However, as an integrated system is a novel concept in vehicular control and energy management.

The objective of this study is to introduce a drive train simulation model considering the impact of trajectory specifications. Also, to maintain the operating point of the engine near its optimal operating region, a fuzzy controller is developed to control a continuously variable transmission (CVT). With respect to fuel economy, the presented model is applied with look-ahead fuzzy controller to manipulate the vehicle speed before going uphill or downhill on a real road. Comparing to the previous studies, the proposed system enjoys from the robustness and fault tolerance of fuzzy systems.

This paper is organized as follows. Section 2 describes the concepts associated with vehicle modeling. Section 3 presents the proposed vehicle model. The control strategy and controller structure are explained in Section 4. Section 5 presents the simulation results and assesses the effectiveness of the proposed model and fuzzy controller. Finally, the concluding remarks are given in Section 6.

Section snippets

Model description

Vehicle is a complex system consisting of many individual components. When the interaction of the vehicle with its environment is considered, it seems that no model can describe precisely the entire behavior of power train. A quasi-static model is presented in this paper to address this issue. Although, the accuracy of a quasi-static model might be generally limited, it can be sufficiently accurate for describing the flow of power in power train. The name of quasi-static is from the stand point

Proposed model

In the proposed model, the drive train is represented by using the backward looking method. The desired speed can be a constant speed as the input of a conventional cruise control or a variable speed. The variable speed can be a drive cycle, an optimizer output or an adaptive cruise controller output (Moon, Moon, & Yi, 2009). We consider that the origin and destination of travel is known in advance and its information can be obtained from GIS. In order to simplify the problem, the road points

Look-ahead fuzzy logic controller

A fuzzy controller, whose rules are extracted based on empirical knowledge of the expert, is developed to manage the energy and to control the transmission ratio. The main objectives of this controller are to implement the look-ahead slopes in order to reduce fuel consumption, and maintain ICE’s speed near its optimal speed range. The advantage of this system is its robustness, since it is tolerant to imprecise measurements and component variations. The fuzzy rules can be easily tuned, if

Simulation and results

A conventional vehicle is modeled in Simulink. This model is presented in Fig. 8. The reference speed is set in adaptive cruise control. Based on the look-ahead slope, a positive or a negative quantity will be added in the reference speed. The total amount is the input of PI controller. The difference between the desired and actual speed is the error signal in the controller input. Hence, the controller should provide a command to eliminate this error. The PI controller and Gain3 block

Conclusions

This paper presented an approach in vehicle modeling based on the road geographical data. From the model, the dynamic position of vehicle is achievable. The applied method uses a combination of the calculated data (distance, altitude and slope) and the static information from GIS. Instantly, the calculated data is compared with the road geographical information (length and altitude). The instantaneous position and slope are extracted. The new information is applied to determine various

Acknowledgment

The authors would like to thank AUTOCRC which provides financial support for this research.

References (18)

There are more references available in the full text version of this article.
View full text