Incremental Evolution of Autonomous Robots for a Complex Task

Author(s):  
Md. Monirul Islam ◽  
S. Terao ◽  
K. Murase
1959 ◽  
Author(s):  
J. S. Kidd ◽  
Robert G. Kinkade
Keyword(s):  

2012 ◽  
Author(s):  
Xiaochen Yuan ◽  
Joseph Shum ◽  
Kimberly Langer ◽  
Mark Hancock ◽  
Jonathan Histon

Author(s):  
Lingtao Huang ◽  
JinSong Yang ◽  
Shui Ni ◽  
Bin Wang ◽  
Hongyan Zhang
Keyword(s):  

2017 ◽  
Vol 12 (1) ◽  
pp. 83-88
Author(s):  
O.V. Darintsev ◽  
A.B. Migranov

In this paper, various variants of decomposition of tasks in a group of robots using cloud computing technologies are considered. The specifics of the field of application (teams of robots) and solved problems are taken into account. In the process of decomposition, the solution of one large problem is divided into a solution of a series of smaller, simpler problems. Three ways of decomposition based on linear distribution, swarm interaction and synthesis of solutions are proposed. The results of experimental verification of the developed decomposition algorithms are presented, the working capacity of methods for planning trajectories in the cloud is shown. The resulting solution is a component of the complex task of building effective teams of robots.


Author(s):  
John Oberdiek

Chapter 2 takes up the complex task of formulating a conception of risk that can meet the twin desiderata of practicality and normativity. Though neither an unreconstructed subjective nor objective account of risk can, on its own, play the role we need it to play in a moral context, the accounts can be combined to take advantage of their respective strengths. Much of the chapter is therefore devoted to explaining how to overcome this recalibrated perspective-indifference. The chapter defends the perspective of a particular interpretation of the reasonable person, well-known from tort law, as a way of bringing determinacy to the characterization of risk. Defending this evidence-relative perspective while criticizing competing belief- and fact-relative perspectives, the chapter argues that it has the resources to meet the twin desiderata of practicality and normativity.


Author(s):  
PAUL A. BOXER

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.


Sign in / Sign up

Export Citation Format

Share Document