Hi, I am working on a project in which I am trying to develop Occupancy grid based on inverse sensor modelling. I am only using it for providing the occupancy of static objects.
I have really basic questions as I am not being able to imagine properly the concepts. I have also read online, attended the course for grid modelling on coursera and also studied some publications but I am looking for simple explanation to explain it to non-technical people.
- What is inverse sensor modelling?
- How to give probability to Grid cells by using Bayesian theorem? Mathematical explanation and calculations or in simple words How to make map through inverse sensor model using bayesian technique (I am using 3 sensors data (Radar,Lidar,Camera) so I am also doing sensor data fusion to predict approximately valid detection)
- How to filter out dynamic detections as I am only interested in static objects detections.
- If possible can someone also refer me some valuable and concrete research publications to get more details.
I will be really thankful for your time and explanations in advance.
Dear M. Asif,
I think you will find many details in answering your questions by jointly checking out these 2 following references:
- T.R. Andriamahefa, " Integer Occupancy Grids: a Probabilistic Multi-Sensor Fusion Framework for Embedded Perception ", PhD Dissertation, 2017 - https://www.theses.fr/2017GREAM010.pdf
- S. Thrun, " Learning Occupancy Grids with Forward Models ", 2001 - http://robots.stanford.edu/papers/thrun.iros01-occmap.pdf
In addition, here are some useful pieces of information:
- Inverse Sensor Modeling is related to Inverse Problem Solving from an general viewpoint,
- Bayesian models are related to stochastic (gaussian) processes and Maximum Likelihood Inference (EM)
- Gaussian processes are related to Radial Basis Functions (RBFs) and kernel machine learning
- Gaussian processes are related to Kalman filtering
- Data Fusion is typically solved by combining Gaussian Processes, (Recursive) Weighted Least Mean Squares and Kalman filtering
Hoping it will be helpful to you,
Regards
You can also check this paper; it was part of my thesis.
Experiments with Simultaneous Environment Mapping
and Multi-Target Tracking
- Badges
- Science topic
- Similar topics
- Engineering
- Electrical Engineering
- Signal Processing
- Image Processing