Matti Pursula

Signal control using fuzzy logic

Applications with fuzzy logic being used in controlling traffic signals have been designed already since the 1970s. The strength of fuzzy logic lies in its capability of simulating the decision-making process of a human, a process that is often difficult to define with traditional mathematical methods. The results of FUSICO-project have indicated that the fuzzy traffic signal control can be the potential control method for signalized intersections.

Saturation Flows at Signal-Group-Controlled Traffic Signals

The main goal of this research was to update the basic saturation flow values of signalized intersections. The secondary goal was to analyze the effects of certain external factors (such as weather, road, and traffic conditions) on saturation flow. The updating is based on extensive field measurements and simulations. Altogether, about 39,000 queues were observed in this study. Field measurements at 30 locations were made according to the method described in the Highway Capacity Manual and simulations were done with the Helsinki University of Technology HUT-SIM simulator, which was calibrated and carefully validated for Finnish road conditions. A summary of calibration parameters is also presented. The new base value for straight-through lanes is 1, 940 vehicles per hour; the previous value was 1, 700 vehicles per hour. In general, the updated saturation flow values of different lane types are 5 to 20 percent larger than the previous base values. The saturation flow models of different lane types are described. The effects of geometric and traffic composition factors, such as percentage of turning vehicles, traffic composition, lane width, and approach grade, were examined and modeled. Effects of weather, road surface, light conditions, and speed level were also analyzed. The drop in saturation flow was about 20 to 30 percent under slippery road and snowy conditions. In rainy conditions, the drop was smaller, about 10 percent. The effect of speed on saturation flow is also described. The most important results of this 2-year project are the saturation flow values for different lane types, knowledge of the effect of external factors (especially during winter), and the large database, which can be used for other purposes. The possibility of using special signal control programs under bad road conditions is discussed. With these kinds of programs, better safety and higher capacity can be achieved.

Mathematical methods on optimization in transportation systems

Preface. Part I: Public Transport Models. 1. Managing and preventing delays in railway traffic by simulation and optimization L. Suhl, et al. 2. Heuristics for scheduling buses and drivers for an ex-urban public transport computing with bus-driver dependencies T. Mellouli, L. Suhl. 3. Computer aided planning of railroad operation T. Siefer, D. Hauptmann. 4. Urban multimodal interchange design methodology R. Garcia, A. Marin. 5. Park-and-Ride station catchment areas in metropolitan Rapid Transit Systems J.A. Mesa, F.A. Ortega. 6. Hub location problems in urban traffic networks S. Nickel, et al. 7. Stochastic assignment to high frequency transit networks: models, algorithms and applications with different preceived cost distributions G.E. Cantarella, A. Vitetta. Part II: General Transport Models. 8. When the MUSICs over. Final results of MUSIC, an EU project to design and implement traffic signal timings which meet a variety of transport goals R. Clegg, et al. 9. Algorithms for the solution of the combined traffic signal optimisation and equilibrium assignment problem M. Maher, Xiaoyan Zhang. 10. Procedures for designing network controls J. Clegg, Yanling Xiang. 11. Approach to congestion optimum toll in traffic networks M.A. Gomez-Suarez, et al. 12. A dynamic network loading model for simulation of pollution phenomena M. DellOrco. 13. Stated preference study of mode choice in the Helsinki metropolitan area J. Kurri, et al. 14. Effects of data accuracy in aggregate travel demand modelscalibration with traffic counts M. Ottomanelli.

Granular Analysis of Traffic Data for Turning Movements Estimation

The paper discusses the principles and the algorithm of granular analysis of data in a specific context of urban traffic monitoring and control (EIS). The proposed granular information processing enables extraction of information on the pattern of journeys from the detailed traffic counts. This facilitates progression from the local optimisation of traffic on individual crossroads to the more holistic optimisation of traffic in a road network. The proposed EIS makes use of readily available stop-line queue data, which is used for adaptive tuning of traffic signals, and adds a data processing layer referred to as granular analysis. It is argued that granular analysis is preferred to statistical data processing since it does not require any assumptions about statistical characterisation of traffic. The granulation algorithm has two distinctive features: (1) the information granules are formed by means of hierarchical optimisation of information density, and (2) the granules are created as hyperboxes thus being readily interpretable in the pattern space. The granular estimates of turning movements are calibrated using an HUTSIM micro-simulator.

MODELING LEVEL-OF-SERVICE FACTORS IN PUBLIC TRANSPORTATION ROUTE CHOICE

A combined revealed preference (RP) and stated preference (SP) survey that was done in the Helsinki, Finland, metropolitan area to reveal the importance of different level-of-service (LOS) factors in public transportation is described. The study was done as a route choice survey between origin-destination zone pairs with two route alternatives that had differences in total travel time, number of transfers, possibility of getting a seat, walking and waiting times, and fare. A separate SP game of transfer environment was also done. The SP study was undertaken among the RP respondents. The levels of the SP attributes were tailored according to the trip in the RP survey. The RP and SP data were analyzed together using structured-tree-type logit models to adjust the variances in different data sets. According to the study results one transfer equals about 10 min of door-to-door travel time, and passengers are willing to travel 15 min longer to get a seat for the trip. The experiences from the use of a tailored postal survey in a SP study were quite positive. However, regardless of study method a very careful definition of the attributes and attribute levels is needed. The results of the survey are quite in line with other similar studies reported in the literature and give an estimate of the passenger response to the public transportation LOS attributes included in the study.

Importance of Fuzzy Sets Definitions for Fuzzy Signal Controllers

The study conducted showed that the definitions of the fuzzy sets associated to the input variables of a given fuzzy signal controller must be taken very carefully as they affect both the actual controller response and the traffic performance.

Studying the importance of fuzzy sets definitions for fuzzy signal controllers

The fuzzy sets adopted for the development of fuzzy signal controllers, and their respective membership functions, are recognised as important aspects related to controller operations. However, the impact of their definitions on traffic signal performance has not been sufficiently explored in the literature. Our aim is to contribute to overcoming this problem. It presents an analysis of the implications of small modifications to some fuzzy sets parameters for fuzzy signal controllers. This analysis focuses upon three aspects. The first looks at modifications to fuzzy signal control tables, generated through MATLAB software and due to variations in the input variables fuzzy sets. The second is the impact of the use of these different control tables on some controller operation output characteristics, as is the green time provided for each controlled intersection approach and cycle length. The last aspect is related to differences observed in traffic performance measures caused by the modifications to the controllers operation characteristics. In order to carry out the analysis of the last two aspects, a simulation study was conducted through HUTSIM simulation software, developed at the Helsinki University of Technology. The studys results show that, in many of the aspects considered, the impact of small modifications to the fuzzy sets parameters is statistically significant at the risk level of 5%

Evaluating off-peak traffic signal control strategies

This article discusses the evaluation of flashing yellow as an off-peak traffic signal control strategy by establishing an interface between a real-time traffic simulation software HUTSIM (Helsinki University of Technology Simulation Model) and a standard NEMA (National Electrical Manufacturers Association) traffic signal controller. The analysis was performed as part of research dealing with evaluation of different off-peak traffic signal operation strategies and their relative impacts on delay, fuel consumption, vehicle emissions, and driver safety. Because of its widespread use, the main emphasis was on the flashing yellow signal control. Delay field study was performed to obtain and analyze real-world data and compare the efficiency of flashing yellow and fixed time control strategies. To widen the research effort and eliminate inaccuracies due to certain assumptions made during the field study, computer simulation was chosen as an effective tool for comparison of different control strategies. Four different strategies were evaluated: fixed time, fully and semiactuated, and flashing yellow. Flashing yellow was found to be the most efficient of the signal strategies. The impact of this type of off-peak signal control on driver safety also was studied, and a summary of results is presented.