A physical model-based analysis of unstructured environments is presented using sonar as a sensing device. Previous methods have relied only on time-of-flight (TOF) methods and have examined only homogenous environments consisting of either smooth or rough surfaces. In this paper, a forward model for the reflection from a class of surfaces with varying degrees of roughness is presented based on the Kirchhoff approximation method. This model integrates different types of environments into a single analytical framework. The echo intensity is parametrized in terms of its energy content and duration, which are functions of the surface roughness, distance, and orientation. The echo-energy and echo-duration maps are introduced to display these parameters. A systematic and robust procedure (ENDURA method) is presented to analyze the reflections and to differentiate and localize the reflecting surfaces. The methodology is verified with experimental results obtained in our laboratory. The results indicate a significant improvement over conventional TOF systems.>

The reflection of wideband pulse signals from rough surfaces is analyzed. A Kirchoff approximation method is employed to describe scattering from rough surfaces, where the usual plane-wave assumption is relaxed to include spherically diverging waves. The signal intensity detected at a point is expressed as an inverse Fourier transform of the product of the frequency correlation function of the channel and the transmitted pulse power spectrum. Coherent and incoherent components of the received signal are determined. The theory is compared with experimental results obtained from the two rough surfaces using the Polaroid transducer operating in air. Good agreement is obtained over the range of angles of incidence from 0 degrees to 60 degrees .>

Conventional time-of-flight (TOF) systems generate sonar maps based on the range estimate obtained from the TOF measurement. This paper describes a novel characterization of sonar reflections from a single sensor based on the physical principles of reflection and scattering from surfaces. Considerable improvement over conventional TOF systems is achieved. Different types of environments are integrated into a single framework. A new parametrization is developed based on the energy content and duration of the echo. These parameters are functions of the surface roughness and orientation. Echo-energy and echo-duration maps are generated that allow the identification of smooth, moderately rough and rough surfaces.>

The physical properties of acoustic sensors are exploited to obtain information about the environment for sonar map building. A theoretical formulation for interpreting the sensor databases on the physical principles of acoustic propagation and reflection is presented. A characterization of the sonar scan that allows the differentiation of planes, corners, and edges in a specular environment is described. A single sensor mounted on an autonomous vehicle in a laboratory verifies the technique. The implications for sonar map building and the limitations of differentiating elements with one sensor are discussed.>