About this Research Topic
In the last years, the computational resources made available by digital boards in conjunction with the continuous advancement of miniaturization technologies and the increase of the operating frequencies have wielded a twofold impact on the evolution of radar systems. On one hand, they allowed to equip such systems with more and more reliable and sophisticated functions relying on complex algorithms that until recently were unthinkable for real-time systems. On the other hand, these factors have contributed to the spread of radar systems in various application contexts of real life such as medicine, defence, communications, manufacturing, automotive, and environment monitoring. In addition, the operating scenarios associated with this multiplicity of applications have become more and more compelling and, more importantly, required the design of suitable algorithms to face with such challenging situations.
The main function of a radar system is the detection of targets competing against unwanted echoes (clutter), the ubiquitous thermal noise, and intentional interference (electronic countermeasures). Once a target is detected, other functions are activated such as the parameter estimation as well as target classification/recognition and tracking. Besides, radar systems can be used to generate high-resolution images of the region of interest for environment monitoring. Clearly, the random nature of the signals in play and the amount of available data entail the exploitation of statistical signal processing techniques that with the advent of the big data era have been included in the general container called Machine Learning. As a matter of fact, this terminology is used to denote all those techniques that learn the underlying model from data and use the latter to make predictions. This task can be accomplished assuming a suitable family of models for data (model-based oriented) or even without any assumption on the underlying models (data-driven oriented).
This Research Topic looks at radar signal processing from the perspective of Machine Learning. More specifically, the interest is on solutions for detection, classification, estimation, and tracking in radar applications and based upon model-based or data-driven machine learning techniques.
We invite investigators to contribute original research papers on the following recommended but not exclusive list of topics:
• Adaptive detection of multiple targets in crowded scenarios
• Clutter covariance matrix estimation and classification techniques
• Tracking algorithms of multiple targets in multi-static configurations
• Radar signal classification
• Electronic Counter-Countermeasures
• Target recognition and classification using HRRP and SAR systems
• Compressive-sensing-based learning techniques
• Polarimetric data classification
• Automotive applications
• Through-the-wall imaging radars
• MIMO radar applications
• Machine-learning-based adaptive radar detection and tracking.
Keywords: Adaptive Detection, Tracking Algorithms, Machine Learning Radar, Estimation, Target Recognition, MIMO Radar, SAR Systems, HRRP Systems
Important Note: All contributions to this Research Topic must be within the scope of the section and journal to which they are submitted, as defined in their mission statements. Frontiers reserves the right to guide an out-of-scope manuscript to a more suitable section or journal at any stage of peer review.