Optimal spectrum sensing over multipath channels

Yifei Huang*, Xiaojing Huang

*Corresponding author for this work

Research output: Chapter in Book/Report/Conference proceedingConference proceeding contributionpeer-review

4 Citations (Scopus)


Wireless propagation phenomena including multipath pose significant challenges to reliable spectrum sensing which is a fundamental requirement for dynamic spectrum access and system coexistence. In this paper, an optimal detection technique along with two reduced-complexity alternatives, modified energy detection (MED) and equal gain detection (EGD), are proposed to improve the detection probability for spectrum sensing over severe multipath channels. By incorporating the resolvable multipaths and multiple receiving antennas into the system model and assuming the availability of a priori temporal correlation about the source signal, these detection methods are derived based on maximum log-likelihood ratio test under low signal-to-noise ratio condition. Simulation results show that the proposed optimal detection outperforms the conventional generalized likelihood ratio test in a multipath environment either with or without a priori information. The proposed MED significantly improves the performance of conventional energy detection after a priori information is exploited. Finally, the proposed EGD performs better than MED and approaches the optimal detection as the number of multipaths increases.

Original languageEnglish
Title of host publication2012 IEEE International Conference on Communications, ICC 2012
Number of pages6
Publication statusPublished - 2012
Event2012 IEEE International Conference on Communications, ICC 2012 - Ottawa, ON, Canada
Duration: 10 Jun 201215 Jun 2012


Other2012 IEEE International Conference on Communications, ICC 2012
CityOttawa, ON


  • cognitive radio
  • maximum likelihood
  • multipath channels
  • Spectrum sensing

Fingerprint Dive into the research topics of 'Optimal spectrum sensing over multipath channels'. Together they form a unique fingerprint.

Cite this