Collision Avoidance in Unstructured Environments for Autonomous Robots: A Behavioural Modelling Approach

2011 ◽  
Vol 403-408 ◽  
pp. 3559-3569
Author(s):  
Chika O. Yinka-Banjo ◽  
Isaac O. Osunmakinde ◽  
Antoine Bagula
Keyword(s):  
Bayesian Network ◽  
Collision Avoidance ◽  
Autonomous Robots ◽  
Reasoning Ability ◽  
Degrees Of Belief ◽  
Bayesian Network Model ◽  
Degree Of Belief ◽  
Unstructured Environment ◽  
Good Potential ◽  
Modelling Approach

Collision avoidance is one of the important safety key operations that needs attention in the navigation system of an autonomous robot. In this paper, a Behavioural Bayesian Network approach is proposed as a collision avoidance strategy for autonomous robots in an unstructured environment with static obstacles. In our approach, an unstructured environment was simulated and the information of the obstacles generated was used to build the Behavioural Bayesian Network Model (BBNM). This model captures uncertainties from the unstructured environment in terms of probabilities, and allows reasoning with the probabilities. This reasoning ability enables autonomous robots to navigate in any unstructured environment with a higher degree of belief that there will be no collision with obstacles. Experimental evaluations of the BBNM show that when the robot navigates in the same unstructured environment where knowledge of the obstacles is captured, there is certainty in the degree of belief that the robot can navigate freely without any collision. When the same model was tested for navigation in a new unstructured environment with uncertainties, the results showed a higher assurance or degrees of belief that the robot will not collide with obstacles. The results of our modelling approach show that Bayesian Networks (BNs) have good potential for guiding the behaviour of robots when avoiding obstacles in any unstructured environment.

LEARNING NAIVE PHYSICS BY VISUAL OBSERVATION: USING QUALITATIVE SPATIAL REPRESENTATIONS AND PROBABILISTIC REASONING

2001 ◽  
Vol 01 (03) ◽  
pp. 273-285 ◽  
Author(s):  
PAUL A. BOXER
Keyword(s):  
Bayesian Network ◽  
Probabilistic Reasoning ◽  
Autonomous Robots ◽  
Visual Observation ◽  
Spatial Representations ◽  
Physical Behavior ◽  
Naive Physics ◽  
Action At A Distance ◽  
General Physical ◽  
Test Scenarios

Autonomous robots are unsuccessful at operating in complex, unconstrained environments. They lack the ability to learn about the physical behavior of different objects through the use of vision. We combine Bayesian networks and qualitative spatial representation to learn general physical behavior by visual observation. We input training scenarios that allow the system to observe and learn normal physical behavior. The position and velocity of the visible objects are represented as qualitative states. Transitions between these states over time are entered as evidence into a Bayesian network. The network provides probabilities of future transitions to produce predictions of future physical behavior. We use test scenarios to determine how well the approach discriminates between normal and abnormal physical behavior and actively predicts future behavior. We examine the ability of the system to learn three naive physical concepts, "no action at a distance", "solidity" and "movement on continuous paths". We conclude that the combination of qualitative spatial representations and Bayesian network techniques is capable of learning these three rules of naive physics.

scholarly journals A Unified View of Causal and Non-causal Feature Selection

2021 ◽  
Vol 15 (4) ◽  
pp. 1-46
Author(s):  
Kui Yu ◽  
Lin Liu ◽  
Jiuyong Li
Keyword(s):  
Feature Selection ◽  
Bayesian Network ◽  
Synthetic Data ◽  
Selection Methods ◽  
Bayesian Network Model ◽  
Real World Data ◽  
Feature Sets ◽  
Unified View ◽  
Optimal Feature ◽  
Different Levels

In this article, we aim to develop a unified view of causal and non-causal feature selection methods. The unified view will fill in the gap in the research of the relation between the two types of methods. Based on the Bayesian network framework and information theory, we first show that causal and non-causal feature selection methods share the same objective. That is to find the Markov blanket of a class attribute, the theoretically optimal feature set for classification. We then examine the assumptions made by causal and non-causal feature selection methods when searching for the optimal feature set, and unify the assumptions by mapping them to the restrictions on the structure of the Bayesian network model of the studied problem. We further analyze in detail how the structural assumptions lead to the different levels of approximations employed by the methods in their search, which then result in the approximations in the feature sets found by the methods with respect to the optimal feature set. With the unified view, we can interpret the output of non-causal methods from a causal perspective and derive the error bounds of both types of methods. Finally, we present practical understanding of the relation between causal and non-causal methods using extensive experiments with synthetic data and various types of real-world data.

scholarly journals Predicting the risk of GenX contamination in private well water using a machine-learned Bayesian network model

2021 ◽  
pp. 125075
Author(s):  
Javad Roostaei ◽  
Sarah Colley ◽  
Riley Mulhern ◽  
Andrew A. May ◽  
Jacqueline MacDonald Gibson
Keyword(s):  
Bayesian Network ◽  
Network Model ◽  
Well Water ◽  
Bayesian Network Model

scholarly journals Impact Mechanism of the Ecological Vulnerability of Highly Developed Islands Based on the Bayesian Network Model—Applied to the Changshan Islands

2021 ◽  
Vol 18 (8) ◽  
pp. 4150
Author(s):  
Keyu Qin ◽  
Haijun Huang ◽  
Jingya Liu ◽  
Liwen Yan ◽  
Yanxia Liu ◽  
...  
Keyword(s):  
Bayesian Network ◽  
Network Model ◽  
Geographical Information ◽  
Global Changes ◽  
Bayesian Network Model ◽  
Island Ecosystems ◽  
Impact Mechanism ◽  
The North ◽  
Ecological Vulnerability ◽  
The Impact

Islands are one of the most sensitive interfaces between global changes and land and sea dynamic effects, with high sensitivity and low stability. Therefore, under the dynamic coupling effect of human activities and frequent natural disasters, the vulnerability of the ecological environment of islands shows the characteristics of complexity and diversity. For the protection of island ecosystems, a system for the assessment of island ecosystems and studies on the mechanism of island ecological vulnerability are highly crucial. In this study, the North and South Changshan Islands of China were selected as the study area. Considering various impact factors of island ecological vulnerability, the geographical information systems (GIS) spatial analysis, field surveys, data sampling were used to evaluate island ecological vulnerability. The Bayesian network model was used to explore the impact mechanism of ecological vulnerability. The results showed that the ecological vulnerability of the North Changshan Island is higher than that of the South Changshan Island. Among all the indicators, the proportion of net primary productivity (NPP) and the steep slope has the strongest correlation with ecological vulnerability. This study can be used as references in the relevant departments to formulate management policies and promote the sustainable development of islands and their surrounding waters

scholarly journals Misper-Bayes: A Bayesian Network Model for Missing Person Investigations

IEEE Access ◽  
2021 ◽  
pp. 1-1
Author(s):  
Denis Reilly ◽  
Mark Taylor ◽  
Paul Fergus ◽  
Carl Chalmers ◽  
Steven Thompson
Keyword(s):  
Bayesian Network ◽  
Network Model ◽  
Bayesian Network Model

Probabilistic Model-Based Fault Diagnosis of the Rotor System

2007 ◽  
Author(s):  
Jiye Shao ◽  
Rixin Wang ◽  
Jingbo Gao ◽  
Minqiang Xu
Keyword(s):  
Fault Diagnosis ◽  
Bayesian Network ◽  
Network Model ◽  
Rotating Machinery ◽  
Rotor System ◽  
Bayesian Network Model ◽  
Diagnosis System ◽  
Model Based ◽  
Core Components ◽  
Fault Diagnosis System

The rotor is one of the most core components of the rotating machinery and its working states directly influence the working states of the whole rotating machinery. There exists much uncertainty in the field of fault diagnosis in the rotor system. This paper analyses the familiar faults of the rotor system and the corresponding faulty symptoms, then establishes the rotor’s Bayesian network model based on above information. A fault diagnosis system based on the Bayesian network model is developed. Using this model, the conditional probability of the fault happening is computed when the observation of the rotor is presented. Thus, the fault reason can be determined by these probabilities. The diagnosis system developed is used to diagnose the actual three faults of the rotor of the rotating machinery and the results prove the efficiency of the method proposed.

Research on Combat Effectiveness Evaluation of Radar EW System Based on Bayesian Network

2011 ◽  
Vol 204-210 ◽  
pp. 1697-1700 ◽  
Author(s):  
Yu Jie Zheng
Keyword(s):  
Bayesian Network ◽  
Evaluation Model ◽  
Effectiveness Evaluation ◽  
Analysis Model ◽  
Key Factors ◽  
Bayesian Network Model ◽  
Simulation Data ◽  
Factors Analysis ◽  
Combat Effectiveness ◽  
Hidden Knowledge

Radar EW system combat effectiveness evaluation is a essential link to Radar system Demonstration, mainly give service to selection, optimization and key factors analysis of Weapon equipment scheme. In this paper, we introduce the Bayesian network model into the area of Radar EW system combat effectiveness evaluation and put forward the concept of combat effectiveness evaluation model based on Bayesian network. The ability to express complex relationship, the ability to express the uncertainty of probability, and the reasoning functions. By learning from Expertise and Simulation data, excavating the hidden knowledge included in both of them, we can build the combat efficiency Analysis model, and then carry out efficient analysis.

Sign in / Sign up

Export Citation Format

Share Document