Artificial immune network-based anti-collision algorithm for dense RFID readers

Abstract Radio frequency identification (RFID) is faced with reader-to-reader collision problem when multiple readers are deployed densely. In scheduling-based methods, the reader-to-reader collision problem can be often mitigated from the viewpoint of optimized resource allocation to readers, which aims at maximizing the total effective interrogation area of an RFID reader network. This paper formulates a reader-to-reader anti-collision model with respect to physical positions, time slots, frequency channels and transmitting power, and thus proposes an artificial immune network with hybrid encoding for resource allocation (AINetHE-RA) to solve this model. In AINetHE-RA, a candidate antibody consists of a location segment, a channel segment and a power segment, where time slots are hidden in the last two segments. According to their respective properties, the location segment and the power segment are encoded by using real number; while the channel segment is encoded by integer number. This is the hybrid encoding format in essence. Correspondingly, in the mutation operator, different mutation strategies are designed for three segments, respectively, which make AINetHE-RA solve this reader-to-reader anti-collision model efficiently. In simulation experiments, the effects of such parameters as time slots, frequency channels, power values and locations are first investigated, and the total effective interrogation area and the number of identified tags are evaluated for the single and multiple density tag distribution. Especially, as an industrial example of non-uniform random tag distribution, the simple sectionalized warehouse management is considered to evaluate the performance of AINetHE-RA. The simulation results demonstrate that the proposed AINetHE-RA algorithm is effective and efficient in mitigating the reader-to-reader collision problem in dense RFID networks and outperforms such methods as GA-RA, PSO-GA and opt-aiNet-RA in finding the optimal resource allocation schemes.