This month´s cover ...

Ayanna Howard and Homayoun Seraji (of JPL) wrote the article "Multi-Sensor Terrain Classification for Safe Spacecraft Landing" that appeared in our Transactions [T-AES vol 40, no 3, pp 1122-1131] in October 2004. A graphical simulation testbed for displaying a spacecraft landing on a planetary terrain is used for evaluation and testing of the multi-sensor information fusion approach. The cover shows the sensor visualization window which displays the sensor data input and corresponding hazard maps deri ved by the algorithm. The hazard map output calculated from the multi-sensor fusion algorithm is displayed at the bottom right-hand comer. This visualization tool can incorporate a wide variety of digital elevation maps (OEMs) that represent different terrains and automatically update current spacecraft dynamic parameters during EDL operations. An additional feature is that data from multiple sensors can easily be incorporated and displayed to the user. The simulation testbed is used to validate and determine the performance of both individual and integrated sensors for hazard assessment and safe site selection. The testbed allows for technology validation under varied environmental conditions, including terrain types, various altitudes, and different lighting conditions. Algorithm performance assessment includes factors such as computational speed, errors in assessment, and comparison with ground truth. For our application, the input parameters to the simulation are: spacecraft conditions (locations/pose and velocity), terrain types (rock field, craters, cliff walls, flat surfaces), and terrain sensors (number of sensors, field-of-view, resolution, noise). The output analyzed includes: selected site safety comparing different sensor suites and site safety as compared with ground truth. [For more on Ayanna Howard, see: IEEE Spectrum, vol 42, no 2, North American Edition, pp 21-22, February 2005.]