Resilient Wireless Sensor Nodes Embedded on Road Reflectors for Vehicle Traffic Monitoring Through Car Engine Sounds
Benjielon D. PascualNorbert U. OngMichael Aldwin S. ReyesCzarina Marie B. Rivera
A Thesis Proposal Submitted to the School of Electrical, Electronics and Computer Engineering in Partial Fulfillment of the Requirements for the Degree
Bachelor of Science in Electronics Engineering
Mapa Institute of TechnologyOctober 2012
Chapter 1INTRODUCTIONThe Philippines is listed as one of the worst in transport related international survey. Manila is the 3rd most worst city for driving according to CNNGo, the travel news website of cable news network or more commonly known as CNN. CNNGo.com added that according to one report, Filipinos perceive traffic congestion as their number one problem. Aside from manila other cities of Metro Manila, being the center of Philippine socio-economic and political activities, has been experiencing heavy traffics and plagued by air pollution because of the large volume of vehicles and vast amount of commuters. Heavy traffic might cause to several problems to the commuters. Some of the problems that might be encountered are delay, fuel consumption and pollution, and road rage. Heavy traffic might also cause car accidents.There are a lot of studies regarding on how to monitor the traffic in highways. Some of the researches are based on 3D video processing technique to monitor the traffic. Here in the Philippines there is an application for Smartphones, MMDA for Android. The said application for Smartphones provide information for the drivers regarding the status of the traffic. The data for this application is gathered through the agencys CCTV camera network and reports from the traffic enforcers. Aside from the traffic monitoring system provided by the MMDA and from the traffic information gathered from the news, there are no other systems or ways to notify and inform the motorists here in the Philippines regarding the traffic. Unfortunately this system is only accessible and applicable for the drivers who can access the Internet within the highways and have Android Smartphones. Applying the concept of wireless sensor nodes, we have thought of a way of monitoring the traffic through the sounds emitted by the vehicles.
The main objective of the study is to build resilient wireless sensor nodes. The study aims to design and create a circuit that detects sound emitted by the engine of moving and stationary vehicles. The study also aims to transmit information from one sensor to another until it reaches the main server by integrating the circuit into a transmitter/receiver. The study also covers to mount the circuit into a traffic reflector found in high ways and to test its resiliency.The primary significance of building resilient wireless sensor nodes is to contribute to the development of future traffic systems. The proposed study will help to inform the drivers ahead of time about the condition of traffic on their way and with this, they can avoid the heavy traffic and take other routes instead. Furthermore, they will also save fuel if there is no heavy traffic. The sensor nodes are mounted on the reflector and it will send the information from one sensor node to another sensor node until it reaches the main server.The proposed study focuses on building resilient wireless sensor for future traffic systems. Resilient Wireless Sensor Nodes are used to detect sounds coming from the engine of the vehicles and to use those sounds to identify the flow of traffic. The study does not cover the classification of vehicles. The device cannot differentiate the sound emitted by the engine of a vehicle from another. We limit our study up to 2 cars. The vibration of the ground produced by the vehicles can affect the resiliency of the circuit.
Chapter 2REVIEW OF RELATED LITERATURE2.1 Possible ways to determine the presence of vehiclesDetermining the presence of a vehicle is one important factor to be considered in this study and there are several technologies that have been developed to detect or sense the presence of these motor vehicles. By far the most common technique is with the use of a CCTV camera. CCTV monitoring systems not only provides security but also prevents people from engaging in criminal or unlawful activities. These are generally used in many cases including real-time monitoring of traffic. As road networks become busier and more congested there is a growing need to monitor the transport network for it is essential for road users to be up-to-date in road occurrences. The automated parking system is another good example in detecting cars. These kinds of system have sensors installed to each parking space that can tell whether or not the space is occupied. Once a car is parked an indicator will mark the bay as occupied and this information is picked up by a Data collector which will be then transmitted and processed in a central data station. The information is released and displayed on a monitor that will tell how many slots are still available. In terms of scientific application for detection of motor vehicles, they generally use the presence of sound. Based on the study SOLAR: sound object localization and retrieval in complex audio environment by D. Hoiem, Y. Ke, and R. Sukthankar, they created this system in order to identify certain types of sound within a complex audio environment. In order for the system to be able to localize and retrieve specific sounds they prefer to identify, they used two classification techniques. The first is with the use of decision trees to select discriminating features and the use of Adaboost to improve classification with an ensemble of trees. In their experiment, they included as object of interest the presence of vehicles through detection of car horns. SOLAR was able to achieve good results on many objects, including the sound of car horns with 46% for 10 FP/hr, 66% for 50FP/hr and 76% for 100FP/hr. FP/hr or false positive per hour is how they measure the datas accuracy. Another technique is the application of inductive loop as shown in Figure 2.1, wherein it works by detecting the change in inductance. A study by Sheik Mohammed Ali, S. involving a Multiple Inductive Loop Vehicle Detection System for Heterogeneous and Lane-Less Traffic was conducted and a sensor was designed which is capable of applying a new measurement scheme for multiple loop system. Its basic feature is to sense and segregate different types of vehicle and to count the number of vehicle in a mixed traffic flow condition. Basically, an inductive loop is simply a continuous coil of wire embedded in the road's surface wherein it enters and exits at the same point. When current first starts flowing in the coil, the coil will build up a magnetic field. The loop resonates at a constant frequency that the detector monitors. When a large metal is detected by the loop such as vehicles, the corresponding inductance and resonance frequency will change. Generally, a compact car will cause a greater increase in frequency than a full size car or truck. Therefore the shift in frequency will help classify and determine the various types of vehicles along the road.
Figure 2.1 Illustration of an inductive loop-based vehicle detection scheme at a junction. 
Table 2.1 Summary of different approaches on detecting vehiclesApproaches on Detecting vehiclesAdvantagesDisadvantage
CCTV cameraRoad users to be up-to-date in road occurrences. Not only provides security but also prevents people from engaging in criminal or unlawful activities.Usually not able to display every square inch of a facility and sometimes causes controversy.
Automated parking systemLess time consuming for drivers to look for a place to park. It would also reduce the traffic congestion caused by those drivers and reduce carbon emissions.Necessitates a maintenance contract with the supplier, higher construction cost per space and redundant systems will result in a higher cost.
Presence of soundHas the ability to measure the speed of vehicles.Other insignificant sounds are always present and must be able to localize sounds created by vehicles
Inductive loopCan help classify and determine the various types of vehicles along the road.
Inability to directly measure speed, requires extensive traffic control and results in congestion and motorist delay and underground wires, are susceptible to being damaged by utility work.
The primary goal of this study is to create a wireless sensor that would help monitor the traffic condition in a specific location. Therefore, the approach that will be used is the detection of sound created by vehicles specifically cars. Sensors will determine the speed of cars through measuring the sound of car engines. The traffic condition will then be established through data processing.
2.2 Different Ways of Detecting the Speed of VehiclesQuickBird Images In year 2009, Wen Liu developed a new method of detecting the speed of vehicle using QuickBird. QuickBird is a satellite that uses Ball Aerospaces Global Imaging System 2000 that captures very high image resolution, which can make houses and buildings visible to our naked eyes from outside of the Earth. He used a pair of QuickBird (QB) panchromatic (PAN) and multi-spectral (MS) images to automatically detect the speed of moving vehicles. This method is tested on different parts of QB images (Figure 2.2 and Figure 2.3). The speed of moving vehicles is detected by analyzing the motion between the images of the PAN and MS in the time lag because PAN and MS sensors have only a 0.2 seconds lag. By using a resolution of 0.6m produced by a PAN image, images of vehicles can be extracted using an object-based approach and with these results, speed of vehicles can be determine. On the other hand, area correlatio