Quantifying detection performance of a passive low-frequency RFID system in an environmental preference chamber for laying hens

Radio Frequency Identification (RFID) systems are widely used in production livestock systems for identification, traceability assurance, and animal registration. An RFID system was implemented and evaluated for use in a four-compartment, two-tier, environmental preference chamber (EPC) to detect hens transiting between compartments. Eight RFID antennae were located near the passages between compartments and each hen wore an RFID tag on one of its legs. The RFID system's performance was assessed by determining detection ranges in a controlled test and by comparing the number of entries into and duration in each compartment as detected by the RFID system to manual assessment of recorded video. The RFID system detection range was revealed to cover a radius of approximately 29 cm from the antenna, excluding a portion of the compartment furthest from the antenna. During a choice-test study, successful detection rates based on duration of stay in compartments were 91.0 +/- 2.6% (mean +/- SD) for trials with groups of birds, and 85.8 +/- 8.0% for trials with individual birds. Based on number of entries into compartments, successful detection rates were 62.6 +/- 11.2% for trials with groups of four birds, and 51.3 +/- 18.4% for trials with individual birds. Sources of entry misdetection included: (i) RFID tag being out of the detection range during sampling; (ii) conflicts caused by multiple RFID tags within the same detection zone; and (iii) visits of duration less than the RFID antenna scan interval. The delay between hen entry and RFID system detection averaged 42.3 +/- 35.7 s and 6.4 +/- 5.2 s in trials with individuals and with groups, respectively. Detection times were within the antenna scanning interval of 15 s (a hen entered a compartment and was immediately detected by the next antenna scan) in 60.3% of the events with individuals and 98.1% of the events with groups. Despite the difference between detected and actual entries (P=0.008 [groups]; P = 0.05 [individuals 1) no significant differences were found between total duration of compartment occupancy over a 24 h test period or average compartment visit duration, as calculated from RFID data and video observations. The RFID system was suitable for characterizing the amount of time birds spent in each EPC compartment, but not reliable for determining number of entries into a compartment. (C) 2015 Published by Elsevier B.V.