Make your own free website on

Data Analysis of Environmental Data for Traffic Control

by Bernhard Krause 1 , Martin Pozybill 2 , and Constantin von Altrock 1 (11/96)
1 ) INFORM GmbH, 2 ) Landesamt für Straßenwesen Baden-Württemberg

Citation Reference: This paper was published at the IEEE International Conference on Fuzzy Logic 1997. The Fuzzy Logic Application Note series is published by Inform Software Corporation on its Internet server to promote the use of fuzzy logic technologies in applications.

Traffic control is based on the analysis of traffic data and environmental conditions. Particularly bad environmental conditions may cause hazard to drivers. In this paper, we discuss the use of fuzzy logic for the analysis of environmental conditions such as road surface condition, visual range and weather conditions detected by local sensor stations and road sensors. Because detection of environmental conditions involves a number of uncertainties, conventional approaches do not deliver satisfactory solutions. The fuzzy logic solution discussed in this paper in contrast:

The discussed fuzzy logic solution was developed for an existing traffic control system of the B27 in Germany, a state highway between Stuttgart city, Stuttgart airport, and Tuebingen. The unit is part of a larger fuzzy logic traffic control system which will be implemented as an additional step in the B27 control system.

1. Traffic Management

The first traffic management systems used in Germany were implemented on roads with frequent accidents caused by fog or icy road conditions. Later, these system were extended to detect and control traffic to increase the traffic capacity. These traffic control systems use several detection stations along the road. These stations employ magnetic sensors for traffic detection, as well as weather stations transmitting environmental data from road surface and the air layer near the ground.

A central traffic control computer collects the data transmitted from the section stations. A control strategy derives an adequate speed limit for every section. The control objectives are:

Along the road of such an "intelligent" highway, alterable road signs posted on traffic sign gantries display speed limits for each lane and display non-regular events such as road work, warnings for traffic back-ups, breakdowns, an accident or dangerous weather conditions (Figure 1) [5].

Environmental and Traffic Sensors Figure 1: Environmental and Traffic Sensors in a Traffic Control System

2. Environmental Data Analysis

The traffic situations depend highly on environmental conditions. An intelligent highway thus should warn drivers of slippery road conditions and low visability. Sensors are used to indicate and classify icy or wet road pavement and to indicate the visual range.

An ideal weather station uses road sensors measuring road surface temperature, road surface moisture, water film depth, and salt content of water film. Near the road, the ideal weather station detects air and ground temperature, amount of precipitation, type of precipitation, wind velocity and direction, sun beam intensity and illumination.

However, most existing weather stations are not equipped with this full range of sensor equipment. In addition, some sensors do not work reliably under all conditions. For example, a standard salinometer needs a wet road surface to measure the remaining salt content on the road. Besides measurement problems, sensors frequently fail because of "biological attacks" such as spiders or butterflies that cover the surface of visual based detectors. Conventional traffic control systems misinterpret this as high rain intensity or low visual range. This can cause completely wrong traffic warnings [3]. Because sensors often stem from different vendors, conventional systems do not use interrelationships between the signals of different detection stations to identify such implausibilities.

3. Fuzzy Logic System Architecture

The fuzzy logic data analysis unit was designed as part of a larger traffic control system. As shown in earlier applications, fuzzy logic is well suited to create solutions for traffic control systems [1, 2]. Figure 2 shows the architecture of the analysis unit, separating a component to verify the sensor information from components to evaluate the road surface and visual range condition.

Architecture of Environmental Data Analysis Figure 2: Architecture of Environmental Data Analysis

4. Sensor Plausibility Analysis

A two-step approach was used to verify the sensor signals. The first step utilizes the fact that no weather signal remains constant. In particular, if the signal jumps abruptly or remains completely constant over time, the sensor signal is considered to be faulty.

In this case, the fuzzy logic system regenerates the information from other sensors. To design this fuzzy logic system, meteorological knowledge about the maximum gradients of all sensor signals, a time frame for a required movement of the signals, and maximum jumps of the gradients to identify discontinuity were acquired from experts.

Road Moisture Fuzzy Logic Module Figure 3: Road Moisture Fuzzy Logic Module

The second step uses four separate fuzzy logic modules to combine interrelated signals:

A. Road Moisture Module

This module, combines all data, that indicates anything about moisture or water on the road. The fuzzy logic module consists of 5 rule blocks (Figure 3) that implement:

  1. A compensation rule block for the hygroscopical behavior of the road moisture sensors. For example, if the salt content is very high, the moisture sensor indicates higher values.
  2. A cross check rule block between detected road surface moisture, the sensors that detect a water film on the road, and the amount of precipitation ions detected during the last 30 minutes. For example, if strong rain was detected over the past minutes, the road must be wet.
  3. A cross check rule block of the humidity sensor using dew point, road temperature, and a moisture sensor. For example, if the dew point is lower than the road temperature and the verified moisture indicates a dry road, the humidity signal must be wrong.
  4. A cross check rule block to the precipitation sensor.
  5. A diagnosis rule block to derive an error message from the given signal situation.

As result, the module produces verified signals of road moisture, amount of precipitation, and humidity.

Road Temperature Fuzzy Logic Module Figure 4: Road Temperature Fuzzy Logic Module

B. Road Temperature Module

This fuzzy logic module contains two rule blocks (Figure 4) to compute:

  1. A verified value of the road surface temperature by cross check of the temperature signal, the gradient of this signal, and the precipitation. For example, the temperature signal can only decrease rapidly if a large amount of precipitation is detected.
  2. A verified value of the freezing point, taking into account salt content and road surface moisture. Because the salt content sensor does not work with dry road conditions, a salt content forecast is used when the signal is not available.

Precipitation Type Fuzzy Logic Module Figure 5: Precipitation Type Fuzzy Logic Module

C. Precipitation Type Module

The verification of the precipitation type is the most complex verification module. This fuzzy logic module verifies existing sensors that indicate the precipitation type by a cross check with the verified signals of road moisture, precipitation quantity, visual range, and other environmental conditions. If the sensor delivers implausible results or is not available, a substitute value is computed. The module consists of a number of rule blocks that:

  1. Indicate if the weather conditions allow for hail or snow. For example, an air temperature level is defined at which snow is implausible.
  2. Compare the visual range with the precipitation quantity.
  3. Aggregate the information with the precipitation quantity and visual range.

D. Visual Range Module

The visual range fuzzy logic module computes a verified value of the visual range by using two rule blocks (Figure 6) that:

  1. Cross check the visual range with the precipitation quantity. For example, there is no fog during heavy rain.
  2. Cross check the visual range with air humidity. For example, fog only occurs during very high humidity.

Visual Range Fuzzy Logic Module Figure 6: Visual Range Fuzzy Logic Module

Using these verified signals, the subsequent two components conclude road and fog conditions. The Road Condition component (Figure 7) uses the qualified values of precipitation type and quantity, as well as the road moisture to indicate a dangerously wet road. Precipitation type, freezing point, and road temperature indicate icy conditions. A final rule block aggregates the information to a standardized road condition classification code. An additional component aggregates the verified values of visual range and precipitation type to compute the standardized visual range classification code.

Analysis of Road Surface Condition Figure 7: Analysis of Road Surface Condition

5. Results

Conventional traffic systems are susceptible to faulty weather sensor signals. The fuzzy logic approach presented delivers more reliable results using meteorological expertise. The solution discussed was implemented using the fuzzy logic development software fuzzy TECH [6]. The table in Figure 8 shows the number of variables, structures, rules and memberships of each component and the complete system used.

Size of Fuzzy Logic Unit And Its Modules Figure 8: Size of Fuzzy Logic Unit And Its Modules

In day to day operation, the fuzzy logic solution has shown that it can prevent traffic control system malfunction in most sensor breakdown situations. In addition "biological attack" situations were detected and misleading rain or fog detection was avoided.

In a complete traffic control system, analysis of environmental conditions is only one component of its functionality. However, faulty weather detection can cause the entire traffic control system to malfunction. Thus, the enhancement of traffic control systems by fuzzy logic greatly improves the reliability of traffic control.

5. Literature

[ 1 ] Kirschfink, H., Schuhmacher, A.: Fuzzy Logik in der Verkehrstechnik (Fuzzy Logic in Traffic Engineering). Proceedings of HEUREKA 1993, pp. 140-160, Karlsruhe 18.-19.3.1993
[ 2 ] Krause, B., von Altrock, C., Pozybill, M.: Intelligent Highway by Fuzzy Logic: Congestion Detection and Traffic Control on Multi-Lane Roads with Variable Road Signs. Proceedings of EUFIT`96, Aachen, Germany, 1996
[ 3 ] Mangold, M., Träger, K., Lindenbach, A.: Wirksamkeit von Streckenbeeinflussungsanlagen unter besonderer Berücksichtigung der Umfelddatenerfassung (Efficacy of traffic control systems focusing on the Detection of Environmental Data), Research Work for the Ministry of Traffic, Kassel 1996
[ 4 ] Zimmermann, H.- J.: Fuzzy Set Theory and Its Applications, Kluwer, Bosten, 1991
[ 5 ] N.N., System Description of B 27 Traffic Management System, Landesamt für Straßenwesen Baden-Württemberg, 1994
[ 6 ] N.N, fuzzy TECH 4.2 User Manual and Reference Manual, INFORM, 1996
[7] fuzzy TECH Home Page at: ""
[8] von Altrock, C., "Fuzzy Logic and NeuroFuzzy Applications Explained", Prentice Hall, ISBN 0-13-368456-2 (1995).