Estimation and Categorization of Errors in Error Recovery Using Task Stratification and Error Classification

Abstract

We proposed an approach to error recovery that uses the concepts of task stratification and error classification. In our method, errors are classified according to the estimated cause into several categories, such as modeling and planning errors. When an error is classified correctly, this increases the probability that the most suitable recovery is performed. In this paper, we describe a procedure for error categorization.

Classification of Causes of Errors in the Human - Robot System

“…One of the most general classifications of robot errors is the classification that has been given in [13]. Nakamura et al all the causes of robot errors classify into four groups: execution errors, planning errors, modeling errors and sensing errors.…”

“…Exposure to vibrations, high acceleration, extremely high and low temperatures, work of the operator in noisy conditions, under low illumination conditions are factors that adversely affect the human -robot system, thus creating the basis for the occurrence of errors. Great changes in temperature and inadequate humidity of the air for the work of robots can cause execution errors [13]. In connection with the aforementioned, Salamani et al [14] have noted that thermal expansion after a cold start has an impact on the occurrence of errors in work of industrial robots.…”

“…Planning errors (also mentioned earlier) can also be classified into a group of errors that originate from an inadequate technological procedure. This error occurs when the interpretation of the system that is used in the computer software does not represent a real system strictly [13]. The cause of the error is mostly a mistake in the setting of a threshold or a parameter during the planning process.…”

“…The cause of the error is mostly a mistake in the setting of a threshold or a parameter during the planning process. Modeling errors occur when the geometry model used in the computer software does not represent the real objects precisely [13]. Since this type of error is directly related to the task execution procedure, it also belongs primarily to the group of errors that originate from the inadequate technological process.…”

“…The fail-safe design is a type of design solution that the consequences of an error are reduced, while it is not necessary that the solution necessarily diminishes the probability of an error. For example, for sensing errors, it becomes necessary to stop the task and make the necessary correction [13]. In this approach, an error can also be automatically compensated by activating additional mechanisms in the event of a malfunction or an error in the system.…”