A plethora of well-established tracking algorithms aim to estimate, over time, the latent kinematic state (e.g., position, velocity, higher order kinematics, or any other spatiotemporal characteristic) of a single or multiple targets based on the available sensory observations, including from several sources. Here, we refer to such techniques as sensor-level trackers. In recent years, there has been a move toward a new paradigm in scene analysis aimed at a higher level understanding of complex

more »

... cenes. The objective may be to circumvent conflict or identify opportunities, automate decision-making and optimize resources allocation as well as future actions for more effective and safer operation of sensing platform(s) or assets. This can entail determining and leveraging, as early as possible, underlying regressors (e.g., objects' intents such as destinations or future actions), social interactions (e.g., hidden group structures and hierarchies), which drive the evolution of a multitarget scene over time, and the capabilities or characteristics of competitors or adversaries. Algorithms that tackle such inference tasks and go beyond kinematic state estimation are dubbed meta-level and adversarial trackers. They belong to a higher system level compared with the sensor-level methods. In this context, adversarial tracking refers to one side estimating over-time the characteristics or strategies of its adversaries and thence calibrating its future behavior, for instance to avoid conflict. Meta-level and adversarial tracking, therefore, present a shift away from the traditional viewpoint of a scene where objects move independently of one another in an unpremeditated manner and without regard to possible competition or group structures, toward an integrated viewpoint where intents, anomalies, group interactions, and characteristics of competitors/adversaries can be automatically learned. This also enables more accurate state estimation by capitalizing on inferred meta-level information. Given the substantial modeling and computational challenges of formulating robust meta-level and adversarial trackers, this Special Section serves to showcase a diverse set of recent relevant technical developments and applications. It comprises of nine selected articles, drawing on recent advances in stochastic modeling, computational methods, statistical filtering, sensing systems, and others. The IEEE TRANSACTIONS ON AEROSPACE AND ELECTRONIC SYSTEMS is an ideal venue for this emerging topic, which calls for a broad high-level systems perspective that can consider decision-making and various modalities or data sources (e.g., radar, optical, RF, acoustics, social media, etc.). The first set of three articles addresses the problem of devising suitable models that describe sophisticated behaviors exhibited by a target and capture the influence of any underlying regressor (e.g., intent, anomaly, etc.). Building on their previous work, Carravetta and White in "Embedded stochastic syntactic processes: A class of stochastic grammars equivalent by embedding to a Markov process" present a theoretical meta-level modeling framework. It is based on the concept of stochastic syntactic processes to relate general grammatical models and processes (e.g., parsing) and statistical estimation problems, most notably by embedding the syntactic process into a Markov random field. This generic formulation is relevant to a wide range of inference problems. Whereas Rezaie et al. propose in "Conditionally Markov modeling and optimal estimation for trajectory with waypoints and destination" a class