Development and Performance Analysis of Integrated Sensing and Communication Systems in Outdoor and Indoor Scenarios

The rapid development of the Internet of Everything (IoE) in sixth-generation (6G) networks has increased the demand for high-rate wireless communication and accurate environmental sensing, motivating integrated sensing and communication (ISAC), in which communication and sensing share spectrum, signals, and hardware resources. This thesis investigates the mutual enhancement between communication and sensing by exploiting existing communication signals for sensing, with two complementary studies addressing outdoor and indoor scenarios. For the outdoor scenario, a cell-free ISAC framework is proposed in which additional downlink (DL) training sequences are employed as dual-functional signals for both communication and sensing. The DL training sequences improve the channel estimation accuracy by approximately 1.5 orders of magnitude. The effects of the DL/UL transmission ratio and pilot proportion are investigated, revealing a saturation region beyond which additional DL or pilot resources provide limited further improvement. To improve the energy efficiency, successive convex approximation (SCA)-based and projected gradient descent (PGD)-based power allocation schemes are developed. At a DL/UL ratio of 8:2, the two schemes reduce power consumption by approximately 17\% and 45\%, respectively, while the average power reductions across pilot proportions are approximately 10\% and 28\%. For the indoor scenario, a reconfigurable intelligent surface (RIS)-aided wideband ISAC framework is developed for fast hierarchical beam searching (HBS). Instead of employing the uplink feedback signal solely for reporting UE indices, the proposed approach further exploits the frequency-dependent spatial signatures induced by beam-squint effect for sensing, while maintaining a fixed RIS phase configuration. Simulation results show that approximately two of the five conventional HBS layers can be skipped with 90\% reliability. The proposed approach achieves the required $0.5^\circ$ direction-estimation root mean square error (RMSE) using only two HBS layers, whereas the conventional HBS requires all five. The impact of fractional bandwidth is also investigated, identifying a turning point around $B_w/f_c=0.2$, beyond which further bandwidth expansion degrades sensing performance because the SNR loss outweighs the additional beam-squint-induced spatial information. In addition, a subcarrier selection algorithm jointly considers subcarrier spatial similarity and received signal strength. With $M=11$ available subcarriers, an average of approximately $7.4$ subcarriers is selected when spatial information diversity is prioritized, reducing overlapping observations while retaining informative sensing measurements. Overall, the results demonstrate that existing communication signals can be further exploited for sensing, enabling enhanced channel estimation and energy efficiency in outdoor cell-free systems and reduced beam training and sensing overhead in indoor RIS-aided systems.