초록 ▼

A motor vehicular collision warning system including a scanned beam sensor, a signal processor, and a vehicle interface system that initiates warnings to the driver or adaptively controls the vehicle. Scene data is used to predict the path of the roadway in front of a subject vehicle. Reflections fr

A motor vehicular collision warning system including a scanned beam sensor, a signal processor, and a vehicle interface system that initiates warnings to the driver or adaptively controls the vehicle. Scene data is used to predict the path of the roadway in front of a subject vehicle. Reflections from manmade and natural scene objects are extracted from the scene data to define a linear feature edge. Once the feature edge is determined, the predicted path of the subject vehicle can be determined using information regarding the vehicle trajectory and the feature edge.

대표청구항 ▼

1. A path prediction method for a vehicular sensing system comprising: (a) scanning a forward field of view by transmitting and receiving radar signals through an oscillating antenna that is driven to oscillate back and forth at a predetermined angle in each direction; (b) detecting a variety of

1. A path prediction method for a vehicular sensing system comprising: (a) scanning a forward field of view by transmitting and receiving radar signals through an oscillating antenna that is driven to oscillate back and forth at a predetermined angle in each direction; (b) detecting a variety of objects in the field of view and generating object data information; (c) extracting object data information which relates to non-vehicle objects; (d) discarding said object data information which pertains to extra large objects not in proximity to likely roadside objects; (e) forming a linear feature edge based upon the extracted non-discarded object data information; (f) extrapolating or interpolating one or more portions of the feature edge when there are discontinuities in the extracted non-discarded object data information; (g) performing one or more smoothing steps upon the feature edge after the feature edge is formed; and (h) determining a predicted path for a vehicle by using an offset with the feature edge. 2. The method of claim 1 wherein smoothing is applied to the feature edge to remove irregularities in the object data information. 3. The method of claim 1 wherein object data information is classified into different categories for determining the object data information to be extracted. 4. The method of claim 1 wherein the offset is a constant and the predicted path is the offset from the feature edge. 5. A sensing system including: (a) an oscillating antenna scanner driven to oscillate back and forth at a predetermined angle in each direction while transmitting and receiving radar signals; and (b) a signal processor which receives object data information and which extracts from such object data information any non-vehicular data information, discards the extracted object data information that pertains to extra large objects not in proximity to likely roadside objects, forms a linear feature edge based upon the extracted non-discarded object data information, extrapolates or interpolates one or more portions of the feature edge when discontinuities exist in the extracted non-discarded object data information, and smoothes the feature edge according to one or more steps, wherein the feature edge is used for predicting the path of a vehicle.