Filter-based stochastic abstractions for constrained planning with limited sensing

2016 ◽  
Author(s):  
Hasan A. Poonawala ◽  
Ufuk Topcu
Keyword(s):  
Limited Sensing

scholarly journals Swarm shedding in networks of self-propelled agents

2021 ◽  
Vol 11 (1) ◽  
Author(s):  
Jason Hindes ◽  
Victoria Edwards ◽  
Klimka Szwaykowska Kasraie ◽  
George Stantchev ◽  
Ira B. Schwartz
Keyword(s):  
Network Topology ◽  
Collective Motion ◽  
Mean Field Theory ◽  
Mean Field ◽  
Interaction Strength ◽  
Central Question ◽  
Agent Interaction ◽  
Sparse Networks ◽  
Swarm Pattern ◽  
Limited Sensing

AbstractUnderstanding swarm pattern formation is of great interest because it occurs naturally in many physical and biological systems, and has artificial applications in robotics. In both natural and engineered swarms, agent communication is typically local and sparse. This is because, over a limited sensing or communication range, the number of interactions an agent has is much smaller than the total possible number. A central question for self-organizing swarms interacting through sparse networks is whether or not collective motion states can emerge where all agents have coherent and stable dynamics. In this work we introduce the phenomenon of swarm shedding in which weakly-connected agents are ejected from stable milling patterns in self-propelled swarming networks with finite-range interactions. We show that swarm shedding can be localized around a few agents, or delocalized, and entail a simultaneous ejection of all agents in a network. Despite the complexity of milling motion in complex networks, we successfully build mean-field theory that accurately predicts both milling state dynamics and shedding transitions. The latter are described in terms of saddle-node bifurcations that depend on the range of communication, the inter-agent interaction strength, and the network topology.

Sensor Optimization for Fault Diagnosis in Multi-Fixture Assembly Systems With Distributed Sensing

1998 ◽  
Vol 122 (1) ◽  
pp. 215-226 ◽  
Author(s):  
A. Khan ◽  
D. Ceglarek
Keyword(s):  
Optimization Procedure ◽  
Hierarchical Optimization ◽  
Dual Purpose ◽  
Distributed Sensor ◽  
Sensor Optimization ◽  
Multi Level ◽  
Sensor Configuration ◽  
Limited Sensing ◽  
Sheet Metal Assembly ◽  
Efficient Distribution

Sensing for the system-wide diagnosis of dimensional faults in multi-fixture sheet metal assembly presents significant issues of complexity due to the number of levels of assembly and the number of possible faults at each level. The traditional allocation of sensing at a single measurement station is no longer sufficient to guarantee adequate fault diagnostic information for the increased parts and levels of a complex assembly system architecture. This creates a need for an efficient distribution of limited sensing resources to multiple measurement locations in assembly. The proposed methodology achieves adequate diagnostic performance by configuring sensing to provide an optimally distinctive signature for each fault in assembly. A multi-level, two-step, hierarchical optimization procedure using problem decomposition, based on assembly structure data derived directly from CAD files, is used to obtain such a novel, distributed sensor configuration. Diagnosability performance is quantified in the form of a defined index, which serves the dual purpose of guiding the optimization and establishing the diagnostic worth of any candidate sensor distribution. Examples, using a multi-fixture layout, are presented to illustrate the methodology. [S1087-1357(00)70801-X]

scholarly journals A Novel Bi-Tuning SSO Algorithm for Optimizing the Budget-Limited Sensing Coverage Problem in Wireless Sensor Networks

2021 ◽  
Vol 11 (21) ◽  
pp. 10197
Author(s):  
Wenbo Zhu ◽  
Chia-Ling Huang ◽  
Wei-Chang Yeh ◽  
Yunzhi Jiang ◽  
Shi-Yi Tan
Keyword(s):  
Wireless Sensor Networks ◽  
Sensor Networks ◽  
Smart Grids ◽  
Wireless Sensor ◽  
Coverage Problem ◽  
Swarm Optimization ◽  
Coverage Ratio ◽  
Sensing Coverage ◽  
Limited Sensing ◽  
Novel Concept

The wireless sensor network (WSN) plays an essential role in various practical smart applications, e.g., smart grids, smart factories, Internet of Things, and smart homes, etc. WSNs are comprised and embedded wireless smart sensors. With advanced developments in wireless sensor networks research, sensors have been rapidly used in various fields. In the meantime, the WSN performance depends on the coverage ratio of the sensors being used. However, the coverage of sensors generally relates to their cost, which usually has a limit. Hence, a new bi-tuning simplified swarm optimization (SSO) is proposed that is based on the SSO to solve such a budget-limited WSN sensing coverage problem to maximize the number of coverage areas to improve the performance of WSNs. The proposed bi-tuning SSO enhances SSO by integrating the novel concept to tune both the SSO parameters and SSO update mechanism simultaneously. The performance and applicability of the proposed bi-tuning SSO using seven different parameter settings are demonstrated through an experiment involving nine WSN tests ranging from 20, 100, to 300 sensors. The proposed bi-tuning SSO outperforms two state-of-the-art algorithms: genetic algorithm (GA) and particle swarm optimization (PSO), and can efficiently accomplish the goals of this work.

scholarly journals Multi-Robot Coverage and Persistent Monitoring in Sensing-Constrained Environments

Robotics ◽  
2020 ◽  
Vol 9 (2) ◽  
pp. 47
Author(s):  
Tauhidul Alam ◽  
Leonardo Bobadilla
Keyword(s):  
Regions Of Interest ◽  
Multiple Robots ◽  
Periodic Trajectories ◽  
Practical Utility ◽  
Physical Experiments ◽  
Constrained Environments ◽  
Limited Sensing ◽  
Contact Sensors ◽  
Multi Robot ◽  
Persistent Monitoring

This article examines the problems of multi-robot coverage and persistent monitoring of regions of interest with limited sensing robots. A group of robots, each equipped with only contact sensors and a clock, execute a simple trajectory by repeatedly moving straight and then bouncing at perimeter boundaries by rotating in place. We introduce an approach by finding a joint trajectory for multiple robots to cover a given environment and generating cycles for the robots to persistently monitor the target regions in the environment. From a given initial configuration, our approach iteratively finds the joint trajectory of all the robots that covers the entire environment. Our approach also computes periodic trajectories of all the robots for monitoring of some regions, where trajectories overlap but do not involve robot-robot collisions. We present experimental results from multiple simulations and physical experiments demonstrating the practical utility of our approach.

scholarly journals Stacked Ensemble of Recurrent Neural Networks for Predicting Turbocharger Remaining Useful Life

2019 ◽  
Vol 10 (1) ◽  
pp. 69 ◽  
Author(s):  
Peyman Sheikholharam Mashhadi ◽  
Sławomir Nowaczyk ◽  
Sepideh Pashami
Keyword(s):  
Short Term Memory ◽  
Remaining Useful Life ◽  
Short Term ◽  
Diagnostic Models ◽  
History Of ◽  
Useful Life ◽  
Long Short Term Memory ◽  
Low Performance ◽  
Limited Sensing

Predictive Maintenance (PM) is a proactive maintenance strategy that tries to minimize a system’s downtime by predicting failures before they happen. It uses data from sensors to measure the component’s state of health and make forecasts about its future degradation. However, existing PM methods typically focus on individual measurements. While it is natural to assume that a history of measurements carries more information than a single one. This paper aims at incorporating such information into PM models. In practice, especially in the automotive domain, diagnostic models have low performance, due to a large amount of noise in the data and limited sensing capability. To address this issue, this paper proposes to use a specific type of ensemble learning known as Stacked Ensemble. The idea is to aggregate predictions of multiple models—consisting of Long Short-Term Memory (LSTM) and Convolutional-LSTM—via a meta model, in order to boost performance. Stacked Ensemble model performs well when its base models are as diverse as possible. To this end, each such model is trained using a specific combination of the following three aspects: feature subsets, past dependency horizon, and model architectures. Experimental results demonstrate benefits of the proposed approach on a case study of heavy-duty truck turbochargers.

Decentralized Platooning With Obstacle Avoidance for Car-Like Vehicles With Limited Sensing

2018 ◽  
Vol 3 (2) ◽  
pp. 835-840 ◽  
Author(s):  
Ioannis M. Delimpaltadakis ◽  
Charalampos P. Bechlioulis ◽  
Kostas J. Kyriakopoulos
Keyword(s):  
Obstacle Avoidance ◽  
Limited Sensing
Sign in / Sign up

Export Citation Format

Share Document