In this paper, we consider the sequential detection problem in a new cognitive radio scenario when the primary user (PU) works with more than one transmit power level. Different from most existing literature where PU is assumed to operate with a constant transmit power only, this new consideration well matches the practical standards, e.g., IEEE 802.11 Series, LTE, LTE-A, and so on, as well as the adaptive powering concept that a user would vary its transmit power under different situations. The targets of the secondary user here are not only to detect the presence of PU but also to recognize PU's transmit power levels. We first formulate a valid sequential detection approach via the modified Neyman-Pearson criterion and then derive the closed-form decision region for each PU's transmit power level. Many interesting discussions are raised. Moreover, the average sample number, a key metric for any sequential detection method, is derived in closed form to facilitate the performance evaluation. The performance comparison of the sequential detection and the fixed sample detection for this multiple primary transmit power levels scenario is analyzed. Finally, the simulation results are presented to verify the correctness of the proposed studies.

In this paper, we consider the sequential detection problem in a new cognitive radio (CR) scenario when the primary user (PU) works with more than one transmit power levels. The targets of the secondary user (SU) is not only to detect the presence of PU but also to recognize PU's transmit power levels. We formulate a valid sequential detection approach via the modified Neyman Pearson (NP) criterion and then derive the closed-form decision region for each PU's transmit power level. Moreover, the average sample number (ASN), a key metric for any sequential detection method, is analytically derived in closed-form to facilitate the performance evaluation. Finally simulation result is presented to verify the correctness of the proposed studies.

In this paper, we consider the sequential spectrum sensing in a practice-matching cognitive raio (CR) scenario, where the primary user (PU) can operate under more than one transmit power levels as regulated in IEEE 802.11 Series, LTE, LTE-A, etc. In this case, the sensing targets not only include detecting the presence of PU but also include recognizing the transmit power level of PU. We propose two different sequential sensing schemes, i.e., detection before recognition and recognition before detection, and also derive the closed form decision regions. Performance of both schemes are compared. Numerical examples are then provided to corroborate the proposed studies.