Novel Indoor Sensor/Access-Point Placement for Full Line-of-Sight (LoS) Coverage

A cost-effective placement strategy of the sensor(s)/access-point(s) for full coverage of the region of interest is always desirable in practice. In this work, we propose a novel systematic approach using the visibility graph of a given arbitrary indoor (simply- or multiply-connected) polygonal environment for full line-of-sight (LoS) coverage where a small (minimum) number of sensors/access points are required. Two difficult but practical challenges, namely, internal obstacles and link-distance restriction, for indoor sensor/access-point placement are also combated by our new approach. Our proposed new approach transforms the sensor/access-point placement problem into a graph partition problem. First, the location(s) of the sensor(s)/access point(s) are initialized by finding all maximal cliques of the aforementioned visibility graph. Then, we propose a new algorithm to eliminate the redundant sensor(s)/access point(s) to further reduce the number of sensors/access points but still retain the full LoS coverage, which significantly reduces the cost without compromising the coverage capacity. According to the simulation results from a number of arbitrary polygons, the number of sensors/access points required by our approach is always less than the theoretical upper bounds derived by this study of the art gallery problem (AGP).