|
An experiment for vehicle guidance system using GPS-GIS integration
These extensions make TransCAD the best tool for working with transportation data. One can use the GIS functions to prepare, visualize, analyze, and present ones work, and use the application modules to solve routing, logistics, and other transportation problems with greater ease and efficiency than any other product. (Refer website 5)
 Fig. 2: All possible paths highlighting shortest path between Powai and Airport Vehicle Guidance Vehicle guidance system consists of algorithms that will display all possible paths between specified origin and destination highlighting shortest path. Also it has algorithms for interpolating, discontinuous GPS data. The methodology adapted for Vehicle guidance can be divided into two parts viz. Algorithm preparation and TransCAD work. Algorithms for vehicle Guidance The algorithms that are prepared for Vehicle guidance are for display of all possible paths between origin and destination and the Interpolating algorithm. The input for the display algorithm is the network file, which contains the road links with their end point positions (White et al., 2000). The output of the algorithm is the file, which contains all the road links between the specified origin and destination. The procedure adapted for this algorithm is that by plotting a graph using the entire link ids and the end points. This is used as a database for the whole algorithm. This graph contains the information about the connectivity of each node. The nearest neighbors for each endpoint are then identified from the graph. The point ids of the specified origin and destination are taken and the nearest neighbors from the origin are identified and from each neighbor its neighbors are called recursively. If the destination is reached the road ids of the path are written to an output file. In this way all the neighbors are called recursively to find all possible paths between origin and destination. If destination is not reached then the particular road links are ignored (Saab, 2000). Another algorithm is prepared for interpolation of positional data that is obtained from GPS. This interpolation is necessary because the GPS data which we get is not continuous because of many reasons viz. obstruction because of tall buildings, non uniform velocity of the vehicle, longer locking time for the GPS receivers in case of Geodetic type of receivers, rapid variations in the acceleration of the vehicle, etc. Hence, a break in tracking is inevitable. In order to get complete and continuous tracking from the GPS data, an interpolating algorithm is prepared to interpolate the values between the break points. The methodology adapted for this algorithm is that first the break in the GPS data is to be identified. For that the difference between two points is calculated. If this difference is more than a specified interval then there is break in the data. Then the midpoint between the two points is calculated and again the difference between the newly obtained point and the original point is calculated. If this is again grater than the given interval then it is again interpolated by calculating their mid point. Also the other side of the midpoint is interpolated. This process in continued until the difference is less than the specified interval. The interval depends on the positional accuracy of the receiver. In this way interpolation algorithm gives continuous GPS data without any breaks. GIS Work The GIS software used for the work is TransCAD. It has all the modules that conventional GIS software should have and also it has several special modules for the application of transportation engineering. TransCAD also has a developer’s kit called GISDK using which one can create his own custom applications to run along with TransCAD (Manual GISDK). GISDK is used for the present problem. Using GISDK a module called Vehicle Guidance is created. Vehicle Guidance has three buttons called Display, Guide and Exit. If Display button is clicked then a dialog box appears which has a pop down menu consisting of different origins. If a particular origin is clicked, then another dialog box appears, hiding the previous one, which contains a pop down menu consisting of different destinations. If a destination is choosen from the list then all possible paths between the origin and destination are calculated using the algorithm. The algorithm gives all the nodes which connect origin to destination. These nodes list is joined with the view and a comparison between the present nodes to the obtained nodes is done and if the node is present then in another field, which is created, a unique value is written. A theme is created with the newly created field and is displayed on the map. This displays all possible paths between the specified origin and destination. In the destination dialog box there is one button called short, which displays the shortest path between the specified origin and destination. The second button created is Guide, which guides the vehicle to the destination. It runs the interpolation program whenever it encounters a break in GPS data. The whole traverse traveled by the vehicle is also shown on the map. The third button is Exit, which exits the program. Also exit buttons are present in each dialog box for instant exiting of the program.  Fig. 3: The path traversed by the vehicle in pilot experiment, after Interpolation Pilot Experiment A pilot experiment is conducted in Mumbai city for testing Vehicle guidance system. Six places in Mumbai city were choosen viz. Powai, Andheri, Airport, Mumbai Central, Chatrapathi Shivaji Terminus (CST) and Gateway of India. These places are taken as Origin and Destination. If any user specifies same origin and destination, then a message is displayed saying that the origin and destination specified are same. If user specifies an origin and destination all possible paths between the two points are calculated and displayed (refer to Fig 1). If the short button is activated, the shortest path between the origin and destination is shown (see Fig 2). Also a vehicle is taken to make a traverse, like a Probe vehicle, with onboard GPS receivers. Three kinds of GPS receivers are used for the purpose viz. Geodetic type TRIMBLE 4000ssi receivers (for details refer website 1) which is of dual frequency type, Mapping type TRIMBLE Pro-XR receivers (for details refer website 1) which is single frequency type, Handheld receivers GPS315 of MAGELLAN (for details refer website 4), which are also of single frequency. Taking all the three receivers on a vehicle, a pilot experiment is conducted by roving on the six places and also on some other places in Mumbai city. A base station was setup before traversing so that the processing mode could be of DGPS mode. After the traverse is completed the data was downloaded from the receivers and as explained earlier, there were breaks between the data collected. This data was then interpolated with the interpolating algorithm and the interpolated co-ordinates are fed to the Vehicle Guidance system. This system then made a line layer using the present data and the line layer was superimposed on the map to get the map of the traversed area (see Fig 3). Conclusions From the data that was obtained from the GPS receivers, a comparison was made between the three receivers for their compatibility with the Vehicle Guidance system. It was observed that Geodetic type receivers had more breaks in the data. The probable reason for this is that the locking time for the antenna is 200 sec in the fastest mode i.e., on-the-fly mode. Hence it requires more time for actually collecting data, thus it is not suitable for vehicle guidance system, even if the positional accuracy of the receiver is more. Mapping type receivers gave a considerable good traverse with less number and length of breaks in comparison with the geodetic type receivers. The positional accuracy of these receivers is not as high as compared with the Geodetic type receivers. They give few meter level accuracy, but it is sufficient for vehicle guidance purposes. Hand held receivers also gave a continuous data with less number and length of breaks but their positional accuracy is in tens of meters, so they are not ideally suited for vehicle guidance purposes. Thus, mapping type single frequency receivers are best suited for vehicle guidance purpose. References
Page 2 of 2
| Previous | |