Abu Dhabi police plans real-time traffic patrol allocation

Abu Dhabi police is making extensive use of geospatial and other modern technologies to automate its processes and make sure that it stays ahead of times to provide better and enhanced citizen services

Emergency services such as police vehicles must provide reasonable service levels in order to ensure public safety. These services are typically provided by vehicles based at fixed locations. The number and placement of vehicles generally influences the quality of services offered. Increasing the number of vehicles is often limited by cost constraints; therefore the efficient deployment of such service vehicles is a crucial issue. Emergency service vehicles must be located in such a way that they may reach any demand point within a maximal response time.

The traffic police in Abu Dhabi perform two major functions: enforcing traffic laws and assisting road users. Within the Abu Dhabi Police department, the Directorate of Traffic and Patrols, which is in charge of all traffic aspects, allocates the traffic patrol vehicles to routine work and to special operations. The special operation vehicles are involved in escorting convoys and motorcades and in enforcing traffic laws and regulations through the deployment of patrol officers in specific areas at certain times, or in dealing with incidents and special events. The routine patrols perform all remaining tasks. At present, the patrol car allocation process is largely manual and relies heavily on the knowledge, experience and expertise of the dispatch operators. The operators divide the Emirate into geographic zones and allocate patrols based on their knowledge about the areas’ needs and other criteria such as traffic flow patterns, population density, etc.  In the event of an accident or call for assistance, the dispatcher calls on one of the patrol vehicles nearby and assigns it the task of dealing with the event.

The current allocation method, while providing acceptable level of service, falls short of providing efficient state-of-the-art utilisation of resources and guaranteeing fast response time in critical situations. Given the advances in Artificial Intelligence and GIS technologies, our ability to monitor traffic flow and the availability of real-time information on traffic conditions, the time is ripe for providing a patrol allocation process based on sound methodology and state-of-the-art technology.

Previously developed systems have shown their operational efficiency and effectiveness by guaranteeing that the police coverage of roads in a specific area is maximised and that the average and maximum response time meet acceptable limits. Most of these systems use specific algorithms in order to determine the configuration of facilities (locations of patrol cars) and assignment of duties to the respective facilities is in accordance with chosen criteria. In most of the allocation applications, the criterion considered is the average response time, measured as the time interval between receipt of a call reporting accident or request for assistance and the arrival of at least one patrol car at the site.

Most of recent research studies and applications concerning patrol cars allocation have produced fully automated systems by using two techniques:  optimal location formulations from the operational research and linear programming fields; and GIS. This project aims to benefit from advances in specific technologies and real-time observation and measurements of traffic conditions in order to develop an efficient real-time patrol allocation system in the Emirate of Abu Dhabi. Such technologies include aspects such as GIS, linear programming, operational research optimization techniques, AI, and trainable expert systems.

 The system receives several inputs (variables) such as: The number of cars which are actively involved in the allocation; Real time traffic conditions (density, flow); Populations density and distribution; Frequency and nature of incidents (accidents, calls for assistance, emergency calls); coverage area; patrol duty parameters (start time, patrol duty cycle, allocation of officer time among various tasks).

A sophisticated algorithm takes into account the input data in real time, the performance constraints and the optimisation criteria in order to produce an optimal allocation strategy for the traffic patrol cars.

The output of the processing system consists of: an initial patrol car allocation strategy for the allocation of traffic patrol cars to geographic zones at the beginning of each patrol duty cycle. This initial allocation is optimised on the basis of the projected traffic pattern and the expected number of accidents and calls for assistance; and a dynamic and real-time allocation strategy for traffic patrol cars that assigns patrol cars to the site of accidents or calls for assistance. The output is determined by the processing algorithm in such a way that optimizes the objective function and meets the stated constraints.

The availability of an efficient patrol allocation system that is based on modern, computer based GIS as well as other advanced technologies, is essential for the smooth flow of the traffic across the roads and highways of the Emirate. In most of the previous research studies, the researchers used GIS data in order to provide static patrol car allocation solution based mathematical formulations using linear programming concepts and GIS. In this research project, the proposed system is able to deal dynamically with varying traffic situations (travel time) caused by accidents or other unexpected events over the space (GIS) and time; and at the same time provide an advanced AI-based decision support service to the officer in charge of monitoring and regulating the flow of traffic under all conditions. By using AI, this system will be more accurate, realistic and comprehensive relative to other systems. In the future, the proposed system will be deployed with other real-time systems which collect real-time data about incidents, traffic, etc. This deployment will enrich the system with the real-time aspect.