Sirf Technology, Inc.

Other

Enuvis, Inc.

Stefano Casadei

Andrew Chou

Benjamin Van Roy

Anant Sahai

Wallace Mann

Jesse Stone

John Tsitsiklis

A method and a system for pattern recognition utilizes an ensemble of reference patterns to represent the possible instances of the models to be recognized; constructs a hierarchy of estimators to simplify and enhance the recognition of the models of interest; approximates complex reference patterns with linear compositions of simpler patterns; fragments complex patterns into local patterns so that interference between the local patterns is sufficiently small for linearization methods to be applicable; constructs estimators during an offline stage to offload calculations from the online signal processing stage; designs model estimators based on optimization principles to enhance performance and to provide performance metrics for the estimated model instances; generates a hierarchy of reference descriptors during the offline stage, which are used for the design and construction of the model estimators. Specific examples are provided for the recognition of image features such as edges and junctions.

Hierarchical method and system for pattern recognition and edge detection

System and method to determine the location of a receiver in a multipath environment are provided. The received signal is correlated with the reference signals associated with the transmitting sources. Each correlation function is processed to derive various types of signal constraints, such as probability densities and uncertainty regions or intervals. In some embodiments, these constraints are for the code-phases and the Doppler frequencies. These signal constraints are transformed into constraints on the receiver variables and then fused together into a unified receiver constraint. A-priori constraints, such as constraints on the location of the receiver or the timestamp, may be incorporated into the unified receiver constraint. Some embodiments estimate a location based also on the estimated Doppler frequency. The constraints used by the invention are based on models of multipath effects and are geared towards mitigating these effects. In one of these models, a probability density for code-phase is obtained by convolving a gaussian distribution with an exponential distribution that describes the extra delay introduced by multipath. Another approach is based on identifying outliers in the set of code-phases. In other approaches, uncertainty region constraints and probability densities are combined. The present invention achieves faster and more sensitive signal acquisition and higher location accuracy in multipath environment, without compromising performance in other environments.

System and method to estimate the location of a receiver in a multi-path environment

Techniques are provided for fine-tuning estimates of a delay value for a sampled signal. One aspect of the invention is to perform, for the sampled signal, coarse-grained calculations of the In Phase and Quadrature (I and Q) correlation integrals at a limited number of points, wherein the calculations are performed over a range of hypothesized delay values. A range of delay values of interest are then determined from the coarse-grained calculations of the I and Q correlation integrals. A subset of I and Q values based on the coarse granularity calculations of the I and Q correlation functions is used to perform a time-domain interpolation to obtain fine-grained values of the I and Q integrals in the range of the delay values of interest. Magnitude calculations are performed based on the fine-grained values of the I and Q integrals. Fine-tuned estimates of delay value are based on the magnitude calculations. Alternatively, fine-tuned estimates of delay value are based on the template-matching approach.

Extracting fine-tuned estimates from correlation functions evaluated at a limited number of values

To determine the clock doppler of a signal receiver, sampled data received from a receiver into are divided into data segments of incremental length. The clock doppler is estimated based on correlating each data segment with the expected signal from each satellite, from a set of satellites, that is overhead the receiver. For each data segment, the correlated result of each satellite is used to refine subsequent calculations of the clock doppler of the next overhead satellite. When the clock doppler calculations for a data segment have been performed using all overhead satellites from the set of satellites, then the results for that data segment are used to refine the calculations for the next data segment. If a current bounds for the clock doppler value is within a pre-determined clock doppler bound-width value and a current bounds for a delay value is within a pre-determined delay bound-width value, then a magnitude template is determined using magnitude calculations of the I and Q correlation integrals obtained at various clock doppler values and various delay values.

Growing community of inventors

Cambridge, MA, United States of America

Stefano Casadei