Combining artificial neural networks and symbolic processing for autonomous robot guidance

Pomerleau, Dean; Gowdy, Jay; Thorpe, C.

doi:10.1016/0952-1976(91)90042-5

Cited by 38 publications

(15 citation statements)

References 1 publication

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The effect of training is to adjust the ways in which the system responds to photosensory information, but that response does not include recognition. We note that Pomerleau et al (1991) also declined to attribute the phrase symbolic knowledge and reasoning to the connectionist driving modules of Navlab, although they used that phrase to characterize the system's use of the annotated map.…”

Section: The Navlab Examplementioning

confidence: 99%

“…The Navlab system (Pomerleau, Gowdy, & Thorpe, 1991), discussed by Vera and Simon, provides a useful example. To illustrate our distinction between symbolic and nonsymbolic information processes, we consider two components of Navlab's robot guidance system that drives a vehicle along roads: an annotated map and a connectionist network for steering.…”

Section: The Navlab Examplementioning

confidence: 99%

See 1 more Smart Citation

Situativity and Symbols: Response to Vera and Simon

1993

View full text Add to dashboard Cite

Vera and Simon (1993) haye provided a helpful and welcome challenge in their articulate questioning of the point of view that emphasizes the situated character of action, including cognition and learning. A full discussion of their arguments and examples requires more space than we are allotted for this response, and we are preparing a longer article that will consider the issues they raised in more detail. We also welcome this opportunity to comment briefly on their provocative arguments. The Issue of SymbolsVera and Simon attribute several beliefs to researchers who are developing situativity theory, some of which we disclaim for ourselves and consider dubious regarding other situativity theorists. However, we do accept their characterization that our view "denies that symbolic processing lies at the heart of intelligence" (pp. 7-8). As we understand the current state of the debate, the issue hinges crucially on the meaning and theoretical status of the concept of symbol.In our view, the emerging scientific practices, empirical findings, and theory that we call situativity theory include the development of ecological psychology (e

show abstract

Section: The Navlab Examplementioning

confidence: 99%

Section: The Navlab Examplementioning

confidence: 99%

Situativity and Symbols: Response to Vera and Simon

1993

View full text Add to dashboard Cite

show abstract

“…In early research into intelligent robots, researchers used genetic algorithm, artifi cial neural network, or combination above two methods to implement robot learning [7][8][9][10][11][12][13][14][15]19]. Wang developed a state generator based on the genetic algorithm, which enables users to generate various motion states without using any reference motion data [16].…”

Section: Literature Reviewmentioning

confidence: 99%

Machine Learning in a Humanoid Intelligent Service Robot

Yen¹,

Hsia

Liao

2014

Journal of Information and Optimization Sciences

View full text Add to dashboard Cite

The service robot described herein is an engineering application combining mechanical technology and computer science. For service robots, intelligence is the core competence. The humanoid intelligent service robot developed at Chienkuo Technology University will possess fi ve intelligent capabilities: marketing, advertisement, acceptance, guidance, and security. It must therefore be taught to obtain intelligence. The purpose of this research is to describe how an instructor teaches the humanoid intelligent service robot. First, according to the required basic action and distinct service script, the robot instructor demonstrates the basic human action. The digital 3D motion capture system captures these motions via video. The movements of the robot instructor's limbs and trunk will be digitized, encoded, and stored in the robot's intelligent computer memory. When service instructions activate the robot, it will replay the motions taught using 27 DC motors, 5 DC servo motors, and 10 solenoids.

show abstract

“…Since only one camera is used, no other sensor data are available, like in [3, 4,5,14] ; and therefore, a 3-D reconstruction is not reliable. Since only one camera is used, no other sensor data are available, like in [3, 4,5,14] ; and therefore, a 3-D reconstruction is not reliable.…”

Section: Ob8tacle Recognition Module (Orm)mentioning

confidence: 99%

<title>Mobile robot navigation with vision-based neural networks</title>

Inigo

Torres

1995

SPIE Proceedings

View full text Add to dashboard Cite

Mobile robot technology is spreading its use in the development of advance manufacturing systems. Methods of multi-sensory fusion data with vision, sonar and limit switches have been developed as the most flexible, but expensive approaches. Other approaches are more common such as buried wire AGV's. They decrease the cost of the mobile robot, but degrade the flexibility of the navigation system as well. This paper uses neural networks (NNs) with only one camera to obtained similar flexibility as the high cost approaches, but in a cost-efficient way.The NNs use translation and perspective information of features in images to determine the proper alignment and position of the mobile robot.

show abstract

Combining artificial neural networks and symbolic processing for autonomous robot guidance

Cited by 38 publications

References 1 publication

Situativity and Symbols: Response to Vera and Simon

Situativity and Symbols: Response to Vera and Simon

Machine Learning in a Humanoid Intelligent Service Robot

<title>Mobile robot navigation with vision-based neural networks</title>

Contact Info

Product

Resources

About