Factor Graph Research Articles

Index Modulation Multiple Access for Unsourced Random Access

In this letter, a novel index modulation multiple access (IMMA) scheme is proposed in the quasi-static MIMO channel. In the proposed scheme, the transmit information bits are divided into two parts, IMMA bits and coded bits. The IMMA bits are employed to select the active compressed sensing codeword transmitted in the first phase, and determine the IMMA pattern. The encoding bits are LDPC encoded, modulated, and transmitted in the second phase through IMMA. In the multiuser detection, the IMMA patterns and active channels are first estimated by approximate message passing (MP) algorithm. Then, the encoding bits are decoded by MP algorithm based on the joint LDPC and IMMA factor graph. Finally, computer simulations are given to demonstrate the performance of the proposed scheme.

A multi-sensor fusion positioning approach for indoor mobile robot using factor graph

Navigation and positioning based on sensor fusion has received great attention in low computational complexity, high positioning accuracy and good robustness for indoor mobile robots. This paper presents a multi-sensor fusion factor graph (MSF-FG) positioning method applying IMU, Odometer and LiDAR sensors. By taking advantage of the sensor characteristics and using factor graph theory, a MSF-FG positioning model is constructed to improve positioning accuracy and reduce computational complexity. In addition, an adaptive function is designed to improve the robustness of the system by dynamically adjusting the weight of each factor. Meanwhile, the proposed algorithm is derived by Gauss-Newton and Levenberg-Marquardt methods. Simulation and experimental results show that compared with the conventional inertial navigation system (INS) and extended Kalman filter (EKF) algorithms, the proposed MSF-FG positioning method not only reduces the mean location error by about 40%, but also reduces the computational complexity and enhances the stability of the system.

Learning spatial-temporal feature with graph product

Earlier works on dynamic spatial-temporal data modelling prefer using spatial-temporal factorized graph convolutional networks (GCNs), which are easy to interpret but fail to capture joint spatial-temporal correlations. Thus, lots of subsequent research focus on constructing a localized adjacent matrix to capture joint features from both spatial and temporal dimension simultaneously. However, their ways of building the adjacent matrices are usually heuristic, which makes the models difficult to interpret. Meanwhile, the lack of theoretical explanations hinders the model’s generalization. We introduce a general framework to model dynamic spatial-temporal graph data from the view of graph product. With the power of graph product, we propose a systematical way of constructing the spatial-temporal adjacent graphs, which can not only improve the model’s interpretability but increase the spatial-temporal receptive field. Under the novel framework, the existing methods can be taken as special cases of our model. Extensive experiments on multiple large-scale real-world datasets, NTU-RGB+D60, NTU-RGB+D120, UAV-Human, PEMS03, PEMS04, PEMS07, and PEMS08, demonstrate that the proposed model can generalize to most of the scenarios with a performance improvement in a significant margin compared to the state-of-the-art methods.

Visual-Inertial-Laser SLAM Based on ORB-SLAM3

At present, visual simultaneous localization and mapping is a hot topic in the field of unmanned systems, which is popular among academic workers because of its advantages of accurate localization, low cost, large amount of information, and wide range of applications, but it still has some problems, including the camera’s vulnerability to the number of feature points and the noise impact of the inertial measurement unit during uniform linear motion. In response to the above problem this paper carries out the research on multi-sensor fusion localization algorithm, the main work is as follows: Based on ORB-SLAM3, a visual-inertial-laser SLAM algorithm is designed. The relative motion of laser location between image frames is obtained from the data of 2D Lidar and laser height sensor. The relative motion of inertial measurement unit between image frames is obtained from inertial measurement unit preintegration. Based on the method of factor graph optimization, the pose of image frame is optimized by reprojection of map point, relative motion increment of inertial measurement unit, and relative motion increment of laser location. The algorithm improves the localization accuracy by about 24.4% over the ORB-SLAM3 visual mode and about 22.6% over the ORB-SLAM3 visual-inertial mode on the data of the UAV physical platform.

LSOS: An FG Position Method Based on Group Phase Ranging Ambiguity Estimation of BeiDou Pseudolite

Due to the influence of indoor space environments, the carrier phase information obtained by the BeiDou pseudo-satellite often has the issue of cycle slips, which makes the user unable to carry out high-precision positioning. Aiming at the problem of ambiguity resolution (AR) and location in large-scale occluded space (LSOS), a factor graph (FG) position method based on group phase ranging ambiguity estimation of BeiDou pseudolite is proposed. Firstly, by introducing the principle of group phase period quantization and utilizing the multi-frequency characteristic of the BeiDou pseudo-satellite, the carrier phase propagation ambiguity of the BeiDou pseudo-satellite can be estimated quickly. On this basis, by introducing the shuffled frog leading algorithm (SFLA) assisted factor graph optimization location estimation method, the BeiDou pseudo-satellite positioning process in LSOS is realized. The experimental results show that the proposed method can solve the problem of fast estimation of ranging ambiguity of BeiDou pseudolite in LSOS, and the ranging accuracy can be improved to two wavelength ranges. In the further location experiment, it is found that the algorithm can not only guarantee the real-time location output but also improve the location precision to sub-meter level under the multi-frequency combination; the optimal location test precision is 9 cm, the maximum positioning error is 50 cm. This method successfully solves the problem wherein the BeiDou pseudo-satellite cannot provide real-time, continuous, and high-precision positioning in LSOS.

Geometric formulations on wireless sensors based spatial protocol for intrusion detection using factor graphs

Restriction on intrusions is one of the foremost aspects of security to conserve the resources in a wireless network. The perimeter of the shadowing zone where in sensing and connectivity are simultaneously changing had to be estimated to prevent intruders. So the proposed work estimates a relationship between the data transmissions and computational time to find the trust factor of channels using factor graphs. Subsequently, estimating the Value of Information (VoI), contention time, and topological significance based on sensing power to alleviate blind spots had been observed to identify intrusions. Factor graph-based routing scenarios capable of handling channel uncertainties in time-triggered wireless sensor communication are indicated with edges and their factors. The process of local interactions with edges provides inference of intrusions to determine global solutions to preserve the network and its data transfer.

Learning Type-2 Fuzzy Logic for Factor Graph Based-Robust Pose Estimation With Multi-Sensor Fusion

Although a wide variety of high-performance state estimation techniques have been introduced recently, the robustness and extension to actual conditions of the estimation systems have been challenging. This paper presents a robust adaptive state estimation framework based on the Type-2 fuzzy inference system and factor graph optimization for autonomous mobile robots. We use the hybrid solution to connect the advantages of the tightly and loosely coupled technique by providing an inertial sensor and other extrinsic sensors such as LiDARs and cameras. In order to tackle the uncertainty input covariance and sensor failures problems, a learnable observation model is introduced by joining the Type-2 FIS and factor graph optimization. In particular, the use of Type-2 Takagi-Sugeno FIS can learn the uncertainty by using particle swarm optimization before adding the observation model to the factor graph. The proposed design consists of four parts: sensor odometry, up-sampling, FIS based-learning observation model, and factor graph-based smoothing. We evaluate our system by using a mobile robot platform equipped with a sensor setup of multiple stereo cameras, an IMU, and a LiDAR sensor. We imitate the LiDAR odometry in structure environments without needing other bulky motion capture systems to learn the observation model of the visual-inertial estimators. The experimental results are deployed in real-world environments to present the accuracy and robustness of the algorithm.

P‐2.7: Slam Fusion Location Mapping Based on Lidar

Aiming at the low accuracy of GNSS/INS integrated navigation caused by the outdoor road environment without GNSS signal or poor GNSS signal, this thesis proposes to combine slam technology with on-board laser scanning system, which can still maintain good robustness in the weak GNSS area. By coupling slam and IMU through the factor graph optimization algorithm, it can improve the mapping results affected by the high speed in the on-board laser scanning system. this thesis uses the system calibration to improve the lio-sam algorithm.Based on the factor graph optimization algorithm, the multi line laser is added to the vehicle lidar scanning system to solve the long-distance and high-precision positioning and attitude determination in the case of weak GNSS signal. The feasibility and effectiveness are verified by the specific application in the field.

Tightly Coupled 3D Lidar Inertial SLAM for Ground Robot

This paper proposes a robotic state estimation and map construction method. The traditional lidar SLAM methods are affected by sensor measurement noise, which causes the estimated trajectory to drift, especially along the altitude direction caused by lidar noise. In this paper, ground parameters in the environment are extracted to construct the ground factors to compress the trajectory estimation drifting along the altitude direction using the characteristics of constant robot pose relative to the ground. Our method uses tightly coupled lidar and inertial to obtain low-drift lidar odometry factors by factor graph optimization. The optimized lidar odometry factors are then added to a global factor graph, together with ground, loop closure, and GPS factors to obtain accurate robot state estimation and mapping after factor graph optimization. The experimental results show that our method has comparable results with advanced lidar SLAM methods, and even performs better in some complex and large-scale environments.

Max-3-Lin Over Non-abelian Groups with Universal Factor Graphs

The factor graph of an instance of a constraint satisfaction problem with n variables and m constraints is the bipartite graph between [m] and [n] describing which variable appears in which constraints. Thus, an instance of a CSP is completely determined by its factor graph and the list of predicates. We show optimal inapproximability of Max-3-LIN over non-Abelian groups (both in the perfect completeness case and in the imperfect completeness case), even when the factor graph is fixed. Previous reductions which proved similar optimal inapproximability results produced factor graphs that were dependent on the input instance. Along the way, we also show that these optimal hardness results hold even when we restrict the linear equations in the Max-3-LIN instances to the form xcdot ycdot z = g, where x, y, z are the variables and g is a group element. We use representation theory and Fourier analysis over non-Abelian groups to analyze the reductions.

A Moving Target Online Location Algorithm Based on Factor Graph

The superior performance of factor graphs compared to Kalman filtering in various fields and the use of factor graph algorithms instead of Kalman filtering algorithms in moving target localization tasks can reduce target localization error by more than 50%. However, the global factor graph algorithm may cause computational delays due to excessive computational effort. A moving target localization algorithm based on a combination of global and incremental optimization with improved factor graphs is proposed to improve localization accuracy and ensure that the computation time can be adapted to the requirements of online location. A reference point is introduced into the incremental calculation process, and it is first determined whether global or incremental calculation is used for this calculation by comparing the distance between the incremental localization results of the calculated reference point. The position of the UAV itself is then corrected by determining the position of the reference point, and this is used to finally locate the target. Simulation results show that the algorithm has good real-time performance compared to the time-consuming global algorithm. The online localization error of moving targets can be reduced by 17% compared to the incremental calculation results of the classical factor graph algorithm.

Indoor Localization Based on Factor Graphs: A Unified Framework

Indoor localization is of pivotal significance for a wide variety of services in the context of the Internet of Things (IoT). Both ranging-based and fingerprint-based localization techniques are promising for employment in harsh indoor environments. Hence, we propose a unified framework based on factor graphs for ubiquitous high-accuracy indoor localization. Our unified framework efficiently integrates ranging and fingerprinting for striking an appealing accuracy versus deployment cost tradeoff, where the crowdsourcing required for the construction of fingerprinting databases can also be addressed with little human intervention. By intrinsically amalgamating the global grid sampling and the regularized importance-resampling techniques, a nonparametric belief propagation algorithm is proposed for achieving the accurate position estimation at the cost of a moderate computational complexity. For improving the robustness to environmental variations, a likelihood-ratio-based approach is employed to detect ranging outliers. Moreover, a low-complexity serial scheduling scheme defined over factor graphs is designed for real-time localization. We design a hybrid ultrawide bandwidth and Wi-Fi localization system relying on off-the-shelf commercial devices and evaluate the proposed unified framework in a typical office building. Our experimental results show that the proposed algorithm outperforms the existing state-of-the-art methods and it is capable of achieving submeter localization accuracy.

Adaptive Adjustment of Factor’s Weight for a Multi-Sensor SLAM

A multi-sensor fusion simultaneous localization and mapping(SLAM) method based on factor graph optimization that can adaptively modify the weight of the graph factor is proposed in this study, to enhance the localization and mapping capability of autonomous robots in tough situations. Firstly, the algorithm fuses multi-lines lidar, monocular camera, and inertial measurement unit(IMU) in the odometry. Second, the factor graph is constructed using lidar and visual odometry as the unary edge and binary edge constraints, respectively, with the motion determined by IMU odometry serving as the primary odometry in the system. Finally, different increments of IMU odometry, lidar odometry and visual odometry are computed as favor factors to realize the adaptive adjustment of the factor’s weight. The suggested method has greater location accuracy and a better mapping effect in complex situations when compared to previous algorithms.

Bayesian detection and tracking of odontocetes in 3D from their echolocation clicks

To study behaviors of marine mammals in a nonintrusive manner, their bio-acoustical signals can be recorded by volumetric hydrophone arrays that provide time difference of arrival (TDOA) measurements for localization and tracking. Multi-target tracking (MTT) in 3-D using TDOA measurements from multiple sensors, however, must cope with non-linear measurement models and high-dimensional states. False alarms, missed detections and unknown data associations impose further challenges, often requiring human operators to annotate the data manually. We propose a data processing chain that automatically detects and tracks odontocetes from their echolocation clicks. The echolocation clicks are detected with a generalized cross-correlation that whitens the instrument noise. Two stages of tracking are performed using a tracking framework based on factor graphs and the sum-product algorithm (SPA). The odontocetes are first tracked in the TDOA domain to remove false alarms and then in the 3-D domain, fusing the tracked TDOAs across all sensors. To efficiently handle the considered non-linear and high-dimensional MTT scenario, particle flow is embedded in the SPA. According to simulation results, the proposed tracking method outperforms the existing approach using manual data annotation. Tracking of Cuvier’s beaked whales (Ziphius cavirostris), whose echolocation clicks are recorded by two volumetric hydrophone arrays, is demonstrated.

LIO-Fusion: Reinforced LiDAR Inertial Odometry by Effective Fusion With GNSS/Relocalization and Wheel Odometry

Reliable state estimation is a prerequisite for autonomous robot navigation in complex environments. In this work, we present LIO-Fusion, a reinforced LiDAR inertial odometry system that optimally fuses GNSS/relocalization and wheel odometry to provide accurate and robust 6-DoF movement estimation under challenging perceptual conditions. LIO-Fusion formulates multi-source sensors fusion based on factor graph, allowing a multitude of relative and absolute measurements which may be degraded, disturbed or even inaccessible. In the LIO-Fusion system, online initialization consists of point cloud feature extraction and matching, IMU preintegration, encoder integration, GNSS calibration and prior-map relocalization. Then, its global reinforcement module detects the reliability of GNSS/relocalization to obtain healthy GNSS/relocalization factors, whereas the local reinforcement module uses a sub-factor graph to fuse prior estimation results for the reinforced local odometry factor. Finally, the basic LiDAR/IMU factors, healthy GNSS/relocalization factors and reinforced local odometry factor are jointly used to constrain the system state in the main factor graph such that low-drift odometry under LiDAR degradation can be reliably obtained and corrected globally. We extensively evaluated the real-time LIO-Fusion system by real-world experiments, and compared its performance to other state-of-the-art methods on large-scale datasets collected in the urban and hazardous environments. Results have shown that LIO-Fusion yielded high precision localization and mapping accuracy as well as robustness to sensor failures.

Performing Bayesian Network Inference Using Amortized Region Approximation with Graph Factorization

Exact inference for large, directed graphical models, also known as Bayesian networks (BNs), can be intractable as the space complexity grows exponentially in the tree-width of the model. Approximate inference, such as generalized belief propagation (GBP), is used instead. GBP treats inference as the Bethe/Kikuchi energy function optimization problem. The solution is found using iterative message passing, which is inefficient and convergent problematic. Recent progress on amortized technique for GBP is an attractive alternative solution that can optimize the Bethe/Kikuchi energy function using (deep) neural networks, requiring no message passing. Despite being efficient, the amortized technique for GBP is applied to undirected graphical models with specific structures and factors, with no guarantee of the approximation quality. This is because the energy function to be amortized is defined by a region (or factor) graph that is ad hoc and difficult to construct to ensure sensible approximations. This paper proposes a new amortized GBP algorithm applied to BN for efficient inference. The proposed algorithm is composed of the following: (i) a new pairwise conversion (PWC) algorithm that converts all the conditional probability distributions in the BN into pairwise factors to facilitate efficient region graph constructions; (ii) following PWC, an improved loop structured region graph (LSRG) algorithm was derived to generate a valid region graph satisfying desired regional properties; and (iii) the energy function defined by the proposed PWC-LSRG region graph can be directly amortized using (deep) neural networks to ensure sensible approximations. Empirical studies show that the proposed amortized PWC-LSRG algorithm is of practical use and significantly improves convergence and efficiency compared to conventional algorithms.

FMGNN: A Method to Predict Compound-Protein Interaction With Pharmacophore Features and Physicochemical Properties of Amino Acids.

Identifying interactions between compounds and proteins is an essential task in drug discovery. To recommend compounds as new drug candidates, applying the computational approaches has a lower cost than conducting the wet-lab experiments. Machine learning-based methods, especially deep learning-based methods, have advantages in learning complex feature interactions between compounds and proteins. However, deep learning models will over-generalize and lead to the problem of predicting less relevant compound-protein pairs when the compound-protein feature interactions are high-dimensional sparse. This problem can be overcome by learning both low-order and high-order feature interactions. In this paper, we propose a novel hybrid model with Factorization Machines and Graph Neural Network called FMGNN to extract the low-order and high-order features, respectively. Then, we design a compound-protein interactions (CPIs) prediction method with pharmacophore features of compound and physicochemical properties of amino acids. The pharmacophore features can ensure that the prediction results much more fit the expectation of biological experiment and the physicochemical properties of amino acids are loaded into the embedding layer to improve the convergence speed and accuracy of protein feature learning. The experimental results on several datasets, especially on an imbalanced large-scale dataset, showed that our proposed method outperforms other existing methods for CPI prediction. The western blot experiment results on wogonin and its candidate target proteins also showed that our proposed method is effective and accurate for finding target proteins. The computer program of implementing the model FMGNN is available at https://github.com/tcygxu2021/FMGNN.

Concatenated Permutation Codes under Chebyshev Distance

Permutation codes are error-correcting codes over symmetric groups. We focus on permutation codes under Chebyshev (ℓ∞) distance. A permutation code invented by Kløve et al. is of length n, size 2n-d and, minimum distance d. We denote the code by φn,d. This code is the largest known code of length n and minimum Chebyshev distance d > n/2 so far, to the best of the authors knowledge. They also devised efficient encoding and hard-decision decoding (HDD) algorithms that outperform the bounded distance decoding. In this paper, we derive a tight upper bound of decoding error probability of HDD. By factor graph formalization, we derive an efficient maximum a-posterior probability decoding algorithm for φn,d. We explore concatenating permutation codes of φn,d=0 with binary outer codes for more robust error correction. A naturally induced pseudo distance over binary outer codes successfully characterizes Chebyshev distance of concatenated permutation codes. Using this distance, we upper-bound the minimum Chebyshev distance of concatenated codes. We discover how to concatenate binary linear codes to achieve the upper bound. We derive the distance distribution of concatenated permutation codes with random outer codes. We demonstrate that the sum-product decoding performance of concatenated codes with outer low-density parity-check codes outperforms conventional schemes.

Refined On-Manifold IMU Preintegration Theory for Factor Graph Optimization Based on Equivalent Rotation Vector

Multisource information fusion is a major area of interest within the field of unmanned technology. In comparison with utilizing the Kalman filter (KF)-based method, the factor graph optimization (FGO)-based method has recently been proposed as an innovative approach. However, the traditional FGO-based inertial measurement unit (IMU) preintegration process is designed for low-accuracy inertial sensors, which leads to insufficient utilization of information and a decrease in positioning accuracy, in the case of tactical-grade or higher accuracy IMUs. Therefore, this article refines the process on the manifold model and utilizes the equivalent rotation vector to accumulate IMU measurements. The main contribution of our work is the refined on-manifold IMU preintegration theory and the application of the multisample algorithm based on the equivalent rotation vector for obtaining more precise preintegrated IMU measurements. The refined on-manifold IMU preintegration theory is applied based on the FGO algorithm in the global navigation satellite system (GNSS)/inertial navigation system (INS) loosely integrated navigation system. The experiments based on simulated and real-world data were conducted and demonstrate that, compared to a high-precision extended KF (EKF) algorithm and optimization methods, the proposed method achieves more accuracy in a dynamic environment.

A Localization Algorithm Based on Global Descriptor and Dynamic Range Search

The map-based localization method is considered an effective supplement to the localization under the GNSS-denied environment. However, since the map is constituted by the dispersed keyframes, it sometimes happens that the initial position of the unmanned ground vehicle (UGV) lies between the map keyframes or is not on the mapping trajectory. In both cases, it will be impossible to precisely estimate the pose of the vehicle directly via the relationship between the current frame and the map keyframes, leading to localization failure. In this regard, we propose a localization algorithm based on the global descriptor and dynamic range search (LA-GDADRS). In specific, we first design a global descriptor shift and rotation invariant image (SRI), which improves the rotation invariance and shift invariance by the methods of coordinates removal and de-centralization. Secondly, we design a global localization algorithm for shift and rotation invariant branch-and-bound scan matching (SRI-BBS). It first leverages SRI to obtain an approximate priori position of the unmanned vehicle and then utilizes the similarity between the current frame SRI and the map keyframes SRI to select a dynamic search range around the priori position. Within the search range, we leverage the branch-and-bound scanning matching algorithm to search for a more precise pose. It solves the problem that global localization tends to fail when the priori position is imprecise. Moreover, we introduce a tightly coupled factor graph model and a HD map engine to achieve real-time position tracking and lane-level localization, respectively. Finally, we complete extensive ablation experiments and comparative experiments to validate our methods on the benchmark dataset (KITTI) and the real application scenarios at the campus. Extensive experimental results demonstrate that our algorithm achieves the performance of mainstream localization algorithms.