Technology Category
Constraint Satisfaction and Optimization: Distributed CSP/OptimizationReasoning under Uncertainty: Relational Probabilistic ModelsIntelligent Robots: State Estimation
Distributed Poisson multi-Bernoulli filtering via generalised covariance intersection
🤖 AI Summary
In distributed multi-object filtering, exact generalized covariance intersection (GCI) fusion of Poisson multi-Bernoulli (PMB) densities remains intractable due to their non-conjugate structure. This paper proposes the first closed-form GCI-based PMB filtering framework. The method constructs an unnormalized product approximation of PMB densities, introduces a Poisson multi-Bernoulli mixture (PMBM) intermediate representation to enable analytical GCI fusion, and designs a closed-loop recursive structure that unifies prediction and update steps for fully distributed filtering. Crucially, it is the first approach to apply GCI rigorously to PMB density fusion—without numerical approximations or sampling—thereby ensuring strict fusion consistency and improved estimation accuracy. Experiments demonstrate superior performance over state-of-the-art distributed PMB and PMBM methods across key metrics: cardinality estimation error, optimal subpattern assignment (OSPA) distance, and label consistency.
Technology Category
Application Category
📝 Abstract
This paper presents the distributed Poisson multi-Bernoulli (PMB) filter based on the generalised covariance intersection (GCI) fusion rule for distributed multi-object filtering. Since the exact GCI fusion of two PMB densities is intractable, we derive a principled approximation. Specifically, we approximate the power of a PMB density as an unnormalised PMB density, which corresponds to an upper bound of the PMB density. Then, the GCI fusion rule corresponds to the normalised product of two unnormalised PMB densities. We show that the result is a Poisson multi-Bernoulli mixture (PMBM), which can be expressed in closed form. Future prediction and update steps in each filter preserve the PMBM form, which can be projected back to a PMB density before the next fusion step. Experimental results show the benefits of this approach compared to other distributed multi-object filters.
Problem
Research questions and friction points this paper is trying to address.
Develop distributed multi-object filtering using PMB and GCI
Approximate intractable GCI fusion of PMB densities
Compare performance with other distributed multi-object filters
Innovation
Methods, ideas, or system contributions that make the work stand out.
Uses generalised covariance intersection fusion rule
Approximates PMB density as unnormalised form
Projects PMBM back to PMB before fusion
🔎 Similar Papers
No similar papers found.
