hjpvandijk / Thesis_ARGoS

0 stars 0 forks source link

Confidence level frontiers #59

Closed hjpvandijk closed 1 week ago

hjpvandijk commented 1 week ago

Adding high (>0.5) probability every time cell is observed as free. Adding low (<0.5) probability every time cell is observed as occupied.

Using L(p) = p/(1-p L(0.5) = 0 = ambiguous.

When hitting an object, there is a chance we are hitting it from an angle, so the observed free boxes hit can actually contain an object. So we decrease the free probability linearly when an object is hit.

Also decreasing distance sensor reading probability with: P = 1 - (10*error)^2 where error is the difference between real and measured distance, based on tests from: https://www.researchgate.net/publication/338041551_Design_and_Realization_of_A_Low_Cost_Clinometer_based_on_ADXL345_Sensor_Ultrasonic_Sensor_and_Android_based_Smartphone