An effective discriminator for differentiating the root causes of packet transmission failures in indoor WSNs

It is well known that there are 2 kinds of causes, namely, channel errors and collisions, which lead to high probability of packet transmission failures in wireless networks. The ability of discriminating the above causes provides many opportunities for implementing high efficient networking protocols in wireless sensor networks. This paper presents EasiPLED, an effective discriminator that can accurately and timely predict these causes. EasiPLED has 3 salient features. First, it investigates frame‐level bit error rate (F‐BER) patterns and statistic characteristics of received signal strength indicator (RSSI) and link quality indicator through extensive indoor experiment studies. The F‐BER is measured at the receiver side by a coarse‐grained method without incurring any overhead. An adaptive RSSI estimator based on error‐based filter is proposed to mitigate effects of environment noises and hardware limitations on RSSI readings. Second, EasiPLED designs an off‐line classifier using machine learning method, which takes a combination of F‐BER and RSSI features as input and outputs the probability of dominant factor that causes transmission failures. Finally, it presents an on‐line discriminator to diagnose the cause of a failed transmission when it occurs at the receiver side, which achieves an accuracy by up to 95.4%. We evaluate the effectiveness of EasiPLED by applying it to link‐layer retransmission scheme and probabilistic polling protocol. Experimental results demonstrate that our method yields a reduction of single‐hop transmission delay by up to 47% and provides high packet delivery ratios as compared to the existing retransmission methods. The EasiPLED‐based probabilistic polling protocol achieves by up to 41.4% throughput gain over the existing method. The paper investigated frame‐level bit error rate patterns and statistic characteristics of received signal strength indicator through extensive indoor experiment studies. The paper designed a classifier using machine learning method, which takes a combination of frame‐level bit error rate and received signal strength indicator features as input and outputs the probability of dominant factor that causes transmission failures, collision, or channel‐error. The paper evaluated the effectiveness of EasiPLED by applying it to link‐layer retransmission scheme and probabilistic polling protocol.