Professional Documents
Culture Documents
Multi-Sensor Data Fusion Techniques For RPAS Detect, Track and Avoid
Multi-Sensor Data Fusion Techniques For RPAS Detect, Track and Avoid
Multi-Sensor Data Fusion Techniques For RPAS Detect, Track and Avoid
net/publication/283298389
Multi-Sensor Data Fusion Techniques for RPAS Detect, Track and Avoid
CITATIONS READS
5 1,336
3 authors, including:
Some of the authors of this publication are also working on these related projects:
Cyber-Physical Systems for Automated Decision Making and Trusted Autonomy View project
Next Generation Air Traffic Management and Avionics Systems View project
All content following this page was uploaded by Roberto Sabatini on 29 October 2015.
Multi-Sensor Data Fusion Techniques for RPAS Detect, Track and Avoid
̂
( ) ([ ] [ ] [ ]) (2)
̂
Page 5 of 12
This is the author pre-publication version. This paper does not include the changes arising from the revision, formatting and publishing process.
The final paper that should be used for referencing is:
F. Cappello, S. Ramasamy, R. Sabatini, “Multi-Sensor Data Fusion Techniques for RPAS Detect, Track and Avoid”, SAE Technical Paper 2015-
01-2475, 2015. DOI: 10.4271/2015-01-2475
∑ ( ̂ )( ̂ ) (14)
This implies that:
( ) ( ̂ ) (3) ∑ ( ̂ )( ̂ ) (15)
In general, the moments appearing in these equations cannot be The resulting expression for the posterior PDF is given by:
computed exactly and so must be approximated. The UT is used for
this purpose. The state prediction is performed by predicting mean ( ) ( ̂ ) (16)
and covariance matrix of which represent the moments of
transformation given by: where:
( ) (4) ̂ ̂ ( ̂ ) (17)
Page 6 of 12
This is the author pre-publication version. This paper does not include the changes arising from the revision, formatting and publishing process.
The final paper that should be used for referencing is:
F. Cappello, S. Ramasamy, R. Sabatini, “Multi-Sensor Data Fusion Techniques for RPAS Detect, Track and Avoid”, SAE Technical Paper 2015-
01-2475, 2015. DOI: 10.4271/2015-01-2475
4. Determine the sigma points and weights to ̂ ∑ (29)
match mean ̂ and covariance matrix .
5. Compute the transformed sigma points ( ), . Sequential Importance Resampling Particle Filter
6. Compute the predicted measurement statistics:
One of the basic challenges in implementation of PF is appropriate
choice of proposal density function. One of the common suboptimal
̂ ∑ (21) choices of proposal density function is transitional prior distribution
[43].
∑ ( ̂ )( ̂ ) (22)
( ) ( ) ( ) (30)
∑ ( ̂ )( ̂ ) (23)
By substituting ( ) into each particle’s weight, a new weighting can
be calculated based on:
7. Compute the posterior mean and covariance matrix
( ) (31)
̂ ̂ ( ̂ ) (24)
In each step of the SIR algorithm, re-sampling is done after state
(25) estimation. After re-sampling, weights of all particles are equal too
where is the number of particles. Hence the above equation is
The Particle Filter modified as follows:
There are different methods for target tracking, but particle filtering ( ) (32)
(PF) is one of the most important and widely used methods. The PF is
implementation of Bayesian recursive filtering using sequential
Monte Carlo sampling method [42]. The PF is used in statistics and
SIR Algorithm
signal processing, which is known as bootstrap filter and Monte
Carlo filter respectively [43]. The target state is represented by Initialise: the particles using prior distribution of target.
where is time. is the set of all observations. The posterior density
function is ( ) is represented by a set of weighted samples. FOR i=1:N
Observations ( ) in the PF method can be non-gaussian. The
main idea of the PF is the estimation of the Probability Distribution a. Prediction: Generate particles using dynamic model
Function (PDF) of the target object by a set of weighted samples
{( ( ) ( ) ) } such that ∑ ( )
[43]. Based ̂ ( ) (33)
on the observations, likelihood of each particle is applied in weight
assignment by the equation below in each step:
b. Weighting: Calculate weight for particle
( ) ( )
( ) (26)
( ) (34)
The PF consists of two stages, prediction and update. In the END FOR
prediction state, next step particles are calculated by using current
step particles and target state transition equations. In the update stage,
Normalize the particle weights by
the weights are assigned to new particles by the following equation
[43]:
̂ ∑
(35)
( ) ( )
(27)
( )
Resampling: Resample the particles and obtain the new part
with same weight and repeat from stage 2.
where is the importance density function which is used in the
sampling function. Posteriori density function is estimated by using
the following equation:
{( )} {( )} (36)
( ) ∑ ( ) (28)
The next step state vector is estimated by using the weighted average
on current step particles with the following equation [43]:
Page 7 of 12
This is the author pre-publication version. This paper does not include the changes arising from the revision, formatting and publishing process.
The final paper that should be used for referencing is:
F. Cappello, S. Ramasamy, R. Sabatini, “Multi-Sensor Data Fusion Techniques for RPAS Detect, Track and Avoid”, SAE Technical Paper 2015-
01-2475, 2015. DOI: 10.4271/2015-01-2475
Instance-Based
Instance-Based Artificial
Artificial Neural
Neural in ∑ [ ] (36)
Learning
Learning Networks
Networks
Err [ ] ( ) (37)
( ) [ ] (38)
Statistical
Statistical Learning
Learning Until some stopping criteria is satisfied
Methods
Methods
Return: Neural-Net-Hypothesis (network).
( ) ( )
Non Parametric Learning Techniques ( ) ⁄
( ) (41)
( ) ( )
Page 9 of 12
This is the author pre-publication version. This paper does not include the changes arising from the revision, formatting and publishing process.
The final paper that should be used for referencing is:
F. Cappello, S. Ramasamy, R. Sabatini, “Multi-Sensor Data Fusion Techniques for RPAS Detect, Track and Avoid”, SAE Technical Paper 2015-
01-2475, 2015. DOI: 10.4271/2015-01-2475
Page 11 of 12
This is the author pre-publication version. This paper does not include the changes arising from the revision, formatting and publishing process.
The final paper that should be used for referencing is:
F. Cappello, S. Ramasamy, R. Sabatini, “Multi-Sensor Data Fusion Techniques for RPAS Detect, Track and Avoid”, SAE Technical Paper 2015-
01-2475, 2015. DOI: 10.4271/2015-01-2475
32. Hall, D.L., “An Introduction to Multisensor Data Fusion,” 47. Welling, M. A, “First Encounter with Machine Learning,”
Proceedings of the IEEE, vol. 85, issue 1, 1997. Irvine, CA.: University of California, 1-93, 2011.
33. Cappello, F., Sabatini, R., Ramasamy, S., Marino, M., “Particle 48. Cosenzo, K. A., Fatkin, L. T., and Branscome, T. A. “Cognitive
Filter based Multi-sensor Data Fusion Techniques for RPAS Uncertainty and Work Shifts in a Real-World Multi-Task
Navigation and Guidance,” Conference: Second International Environment (No. ARL-TR-3515),” Army Research Lab
Workshop on Metrology for Aerospace (MetroAeroSpace Aberdeen proving Ground MD: Human Research and
2015), At Benevento, Italy, 2015. Engineering Directorate, 2005.
34. Wan, Eric A., and Van Der Merwe, R., “The Unscented Kalman 49. Shahbazian, E., Rogova, G., and DeWeert, M. J. “Harbour
Filter for Nonlinear Estimation”, Adaptive Systems for Signal Protection through Data Fusion Technologies,” Springer Science
Processing, Communications, and Control Symposium 2000. & Business Media, 2009.
AS-SPCC. The IEEE 2000. IEEE, 2000. 50. Konopko, K., and Janczak, D., “Maneuver Detection Algorithm
35. Van Der Merwe, R., “Sigma-point Kalman Filters for Based on Probability Density Function Estimation with Use of
Probabilistic Inference in Dynamic State-space Models”, RBF and HRBF Neural Network,” In Photonics Applications in
Dissertation, Oregon Health & Science University, 2004. Astronomy, Communications, Industry, and High-Energy
36. Oh, S-M., “Nonlinear Estimation for Vision-Based Air-to-Air Physics Experiments II (pp. 625-629). International Society for
Tracking”, Diss. Georgia Institute of Technology, 2007. Optics and Photonics, 2004.
37. Challa, S., “Fundamentals of Object Tracking”, Cambridge 51. Cuadrado-Gallego, J.J., Braungarten, R., Dumke, R.R., Abran,
University Press, 2011. A. (Eds.), “Software Process and Product Measurement
38. Doucet, A., & Johansen, A. M. (2009). “A Tutorial on Particle International Conference,” IWSM–MENSURA 2007, Palma de
Filtering and Smoothing: Fifteen Years Later”, Handbook of Mallorca, Spain, November 5-8, 2007.
Nonlinear Filtering, 12, 656-704. 52. Demmel, S., Gruyer, D., & Rakotonirainy, A., “V2V/V2I
39. Hussain, M.S., “Real-Coded Genetic Algorithm Particle Filters Augmented Maps: State-Of-The-Art and Contribution to Real-
for High-Dimensional State Spaces”, (Doctoral dissertation, Time Crash Risk Assessment,” In Proceedings of 20th Canadian
UCL (University College London)), 2014. Multidisciplinary Road Safety Conference, The Canadian
40. Lai, C.K., Lone, M., Thomas, P., Whidborne, J., Cooke, A., Association of Road Safety Professionals, Ontario, Canada,
“On-Board Trajectory Generation for Collision Avoidance in 2010.
Unmanned Aerial Vehicles”, Proceedings of the IEEE 53. Dontabhaktuni Jayakumar, D. P., & Arunakumari, D. “Satellite
Aerospace Conference, pp. 1-14, 2011. DOI: Image Fusion in Various Domains,” International Journal of
10.1109/AERO.2011.5747526 Advanced Engineering, Management and Science (IJAEMS),
41. Angelov, P., “Sense and Avoid in UAS: Research and vol. 1, issue 3, June 2015.
Applications,” John Wiley & Sons, 2012. 54. Garzelli, A., Aiazzi, B., Baronti, S., Selva, M., and Alparone, L.
42. Cheng, Q. and Bondon, P., “A New Sampling Method In “Hyperspectral Image Fusion,” In Proceedings of the
Particle Filter,” 17th European Signal Processing Conference Hyperspectral 2010 Workshop (pp. 17-19), 2010.
(EUSIPCO), 2009. 55. Pramanik, S., & Bhattacharjee, D. “Multi-sensor Image Fusion
43. Parseh, M-J., Pashazadeh. S., and Seyedarabi, H., “Improved Based on Moment Calculation,” 2nd IEEE International
Particle Filtering Algorithm for Maneuvering Object Tracking Conference on Parallel Distributed and Grid Computing
Using Deformation Detection,” 2nd International Conference on (PDGC), pp. 447-451, December 2012.
Computer and Knowledge Engineering (ICCKE), October 18- 56. Toet, A., “Cognitive Image Fusion and Assessment,” INTECH
19, 2012. Open Access Publisher, 2011.
44. Mohri, M., Rostamizadeh, A., and Talwalkar, A., “Foundations 57. Toet, A., & Hogervorst, M., “Real-time Full Color Multiband
of Machine Learning,” The MIT Press, ISBN: 9780262018258, Night Vision,” INTECH Open Access Publisher, 2010.
2012. 58. Stathaki, T. “Image Fusion: Algorithms and Applications,”
45. Bengio, Y., Courville, A., and Vincent, P., “Representation Academic Press, 2011.
Learning: A Review and New Perspectives,” IEEE Trans Pattern
Anal Mach Intell. 2013 Aug; vol. 35, issue 8, pp. 1798-828,
DOI: 10.1109/TPAMI.2013.50.
46. Anthony, M. H. G., and Biggs, N., “Computational Learning
Theory,” Cambridge University Press, 1997.
Page 12 of 12