Some users of conversational systems sometimes complain of difficulties of understanding the operational sequences. The reason of the difficulties seems to be attributed that the users can not proceed through “plan-do-see” cycles smooth, because some failures of their system interface design exist and thus the system models are unclear. In order to evaluate the user interface conditions which enable their users to proceed with “PDS” cycles smooth, this study proposes Plan-Do-See (PDS) Analysis which can be used to evaluate system specifications or trial manufactures, without conducting actual user tests. The proposed PDS Analysis is tested to evaluate the usability of a personal facsimile, before conducting its user test. The difficulties of its operation obtained by the user test were almost forcasted by PDS Analysis, however, several usability problems related to user's mental models of operational sequence could not be forcasted. PDS Analysis is, therefore, effective in evaluating usability but is of limits for some kinds of usability problems.

View full abstract

In a task to keep water temperature constant by manual dial operation, the correlations between feedback informations given to subjects and their control behavior were investigated. In this study, different feedback informations were given to the subjects who are divided into four groups. That is, informations concerning water temperature, displayed either in analog or digital form, were shown spontaneously or with a two minute delay. Each subject performed fifteen trials. Deviations from target temperature, number of operations and frequency which the subject needed the temperature information were analyzed.
As a result the following two results were obtained. With increasing learning experience, the information content which the subjects needed changes from the dynamic characteristics to the current conditions of the system. The kind of feedback information such as analog or digital form influences dial operation.

View full abstract

The purpose of this study is to develop the inferring method of perceived fatigue with physiological indices using a neural network (NN). Seven healthy male subjects performed five intermittent handgrips, and one continuous handgrip at 10% maximum voluntary contraction for 10 minutes. The physiological indices and the perceived fatigue were measured during the last 2 minutes of both handgrip patterns. The structure of NNs was following: (1) input layer had 7 units (physiological indices), (2) output layer had 1 unit (perceived fatigue), and (3) one or 2 hidden layers in which those layers had different units. Every NN was continuously learned with learning-data until total errors became less than 0.3, 0.2 and 0.1. Thereafter, unlearning-data were put into learned NN, then mean error was computed. The learned NN showed the smallest mean error in which total error used was less than 0.2 with the structure of 7×6×5×1. It was concluded that NN could be used for the inference of perceived fatigue under the extent of an error by the input of the physiological indices.

View full abstract

In speech recognition systems speakers may consider their speech as an input to recognition devices. However, from the view of human-machine interaction, human vocal behaviors may also be an output affected by the performance of the recognition system.
Subjects were requested to speak out color names of rectangles painted with either blue, green, red, white, or yellow on a CRT. A part of the vocal responses of the subjects in the experimental condition were intentionally misrecognized and rejected after a series of successful recognition. The subjects' vocal responses became significantly louder and longer after misrecognition compared with the voice emitted just before the rejection. Female subjects altered their speech style more promptly than males did. Male subjects changed their vocal fundamental frequency. The subjects in the control group, where it is so designed that speech was always successfully recognized, did not show specific loudness changes.

View full abstract

Differences in the determined levels, fluctuations and measurement time of critical flicker fusion frequency were studied experimentally by comparing the results for unskilled and skilled subjects, as well as the measuring methods for determining fusion frequency limits and referred adjustments. Three methods were used for adjustments, i. e., two-way adjustment (either linear or non-linear) and one-way adjustment. The experiments consisted of Experiment I, II, III-1 and III-2. The results of Expriment I with 18 unskilled subjects produced a CFF level of 36.12Hz (requiring 9.37s for the measurement) for the two-way adjustment method (linear), 35.13Hz (10.57s) for the one-way adjustment method, and 32.23Hz (12.75s) for the limits method. Highly significant differences were found in mean values of iso-variances between the adjustment and limits methods. Regarding measurement time, highly significant differences were found in both variances and mean values. The results of Experiment II with 9 skilled subjects produced 38.27Hz (12.63s) with the two-way adjustment method, 37.64Hz (13.85s) with the one-way adjustment method, and 34.90Hz (13.11s) with the limits method. There were no differences in measurement time. As in Experiment I, highly significant differences were found in CFF with Experiment II. These results led to the conclusion that the two-way adjustment method would be better. Experiment III-1 and III-2 were conducted by making use of the variance rate chosen for the two-way non-linear adjustment method and by allowing fine adjustments near CFF (35Hz). As a result, CFF levels of 35.82Hz (11.61s) with 13 unskilled subjects and 37.79Hz (14.11s) with 10 skilled subjects were obtained.

View full abstract

All the human factors requirements for manned space system development have not been quantitatively specified. Therefore, human factors engineers often have difficulty evaluating the system validity with quantitative measures. In addition, the human factors engineers should consider zero gravity in the evaluation process because human performance in zero gravity, specifically motor activities, is quite different from that of on the earth. This report specifies the following points that human factors engineers should consider when evaluating manned space systems with quantitative measures and considering zero gravity. (1) Crew working posture in zero gravity, (2) Processes taken in the system evaluation, (3) Techniques for simulating zero gravity on the earth, (4) Measures for quantitative evaluation, (5) Factors affecting the crew performance on orbit.

View full abstract