A Minimalist Path Detection Approach for Wireless Sensor Networks

Most object tracking techniques find the exact locations of an object, while many other applications are only interested in the object's path rather than its locations. This provides an opportunity to reduce the existing tracking methods' resource usage by only focusing on the path detection. Taking advantage of this opportunity, this article presents two new minimalist approaches for accurately detecting unknown available passages in a sensor field without requiring the exact locations of the objects. In the first approach, each sensor sends its own location to a base station when it senses an object of interest. The base station uses b-spline curves to build the object's path online. Since each sensor sends its location data just once per new path, the first approach is a minimalist approach. The second approach is offline and uses non-uniform rational b-spline (NURBS) curves. Since NURBS needs weighted locations, each sensor sends its own location in addition to the number of times it has sensed an object based on the object's weight. Using the same simulation models, both approaches greatly reduce power consumption and improve the accuracy of the computed paths. The NURBS approach has proved to be robust on false alarms and improves the accuracy of path detection up to 95 percent, which is very close to detecting the object's actual path.