Recognition-Based Autonomous Navigation Systems, Methods, and Non-Transitory Computer-Readable Media

A recognition-based autonomous navigation method, comprising:

• receiving sensor data associated with an environment of an autonomous or semi-autonomous platform;
• deriving at least one recognition metric from at least one sensed, computed, inferred, or estimated value;
• normalizing the at least one recognition metric to generate at least one normalized recognition metric;
• generating a bounded recognition factor from the at least one normalized recognition metric;
• generating a fused perceptual output by applying the bounded recognition factor to recognition-adjusted weighting of a plurality of sensor-derived inputs;
• generating a detected object output by applying the bounded recognition factor to at least one neural inference value, activation value, feature value, or detection-related quantity;
• generating a path or trajectory by applying the bounded recognition factor to at least one path-planning cost;
• and
• determining execution priority for at least one task, action, or control decision by applying the bounded recognition factor to at least one risk-related input.

A non-transitory computer-readable medium storing instructions that, when executed by one or more processors, cause the one or more processors to:

• receive sensor data associated with an environment of an autonomous or semi-autonomous platform;
• derive at least one recognition metric from at least one sensed, computed, inferred, or estimated value;
• normalize the at least one recognition metric to generate at least one normalized recognition metric;
• generate a bounded recognition factor from the at least one normalized recognition metric;
• apply the bounded recognition factor to recognition-adjusted weighting of a plurality of sensor-derived inputs to generate a fused perceptual output;
• apply the bounded recognition factor to at least one neural inference value, activation value, feature value, or detection-related quantity to generate a detected object output;
• apply the bounded recognition factor to at least one path-planning cost to generate a path-planning output;
• and
• apply the bounded recognition factor to at least one risk-related input to generate a recognition-based priority for at least one task, action, control decision, or resource allocation.