Wide Area Search Munitions (WASMs) are unmanned aerial vehicles with an array of onboard sensors, a warhead or other kill mechanism, and autonomous flight capabilities. WASMs have played many important roles in the modern battle field including reconnaissance, search, battle damage assessment, or communications relay. The ATR (automatic target recognition) system in WASM is used to identify potential targets and broadcast this information to other WASMs. However, target identification is subject to errors. For example, two WASMs might detect the same target, but associate with that target two separate target tracks. Conversely, two WASMs might detect separate targets, but assign the same target ID to both targets. We call this problem as the object misidentification (OMI) problem. To solve the OMI problem for a set of points in 4-dimensional space (time, longitude, latitude and altitude), we will investigate data mining techniques like clustering to identify the individual route of each target (detected point). The overall goal is to reconstruct the battle space using statistical, optimization, and data mining approaches.
Week 1 and 2
Multiple hypothesis tracking for multiple target tracking