Abstract: In Internet of thing, short-range communication is a common communication system, where a sensor node will collect signals from multiple physical nodes. Because the system uses a wireless shared channel, signal transmissions together from multiple sources will cause collision. Usually, the collision is resolved by a random multiple access method, but the communication efficiency is not high. Signal separation and coding technology can directly recover the source from the collision signals, so it can significantly improve the communication efficiency. However, the method requires predicting or estimating signal fading coefficients to ensure good performance. This paper proposes a Viterbi decoding method with finite symbol separation, which can accurately recover the collision signal without the fading coefficients, through estimating the dictionary matrix. In the experiment, we establish two ultra-high frequency short-range wireless communication system based on FM0 code, using simulation and software radio, respectively. The experimental results show that the throughput of the proposed method arrives at about 0.61, which improves the traditional ALOHA system by bout 0.25.