LLM-Driven Cognitive SemCom Systems

With the rapid developments in artificial intelligence (AI) and communication technologies, mobile robots, particularly unmanned aerial vehicles (UAVs), have seen increasingly widespread applications across various domains, playing an important role in the filed of communication services. For instance, in the real-world search and rescue scenarios, the UAVs can quickly access remote or hazardous areas that are difficult for humans to reach, providing timely and critical information to support rescue operations. However, in UAV image communication, wireless communication process faces further constraints when the images captured by UAVs need to be transmitted to the receiving end. To be specific, with the arrival of the era of everything intelligent and connected, the high-quality spectrum resources are decreasing while the data volume of images and video streams increases, which poses serious challenges to existing communication systems to achieve low latency and high data transmission rates within limited spectrum resources. These challenges highlight a fundamental conflict between spectrum resource limitations and the demand for high-quality services. To address this problem, researchers are increasingly exploring SemCom, a new paradigm of the next-generation communication system that transcends bit-level accuracy to prioritize the transmission of meaningful information.