The Characteristic Of Multiple-object Visual Tracking In College Students And Its Enlightenment For Application

As mobile intelligent robots are used in an increasingly wide range of application scenarios,the study of multi-object tracking techniques has become extremely important.This is because intelligent robots,especially mobile robots,are often limited in the type and number of sensors they can carry,and methods to achieve multi-object tracking with limited sensor resources need to be sought.To address these limitations and challenges,researchers try to explore a variety of approaches to achieve multi-object tracking for intelligent robots.However,the patterns and characteristics of human behaviour observed in existing psychological multi-object tracking studies are not applicable to the optimisation of intelligent robots.This is because the typical multi-object tracking paradigm mostly examines individuals tracking objects while discriminating between targets and distractors with very similar characteristics,which consumes a large number of cognitive resources and significantly reduces the number of objects that individuals can track in real time with limited resources.But real-life multi-object tracking tasks often do not require careful marking and discrimination between targets and distractors.Therefore,to design an intelligent robot system based on the results of traditional multi-object tracking studies,marking and distinguishing between targets and distractors while tracking would only reduce the adaptability of the robot system and increase consumption.To address these problems,this study proposes a new variant of multi-object tracking that no longer distinguishes between distractors and targets in the tracking task,and explores the multi-object tracking capacity,tracking effects,and tracking strategies of university students through three studies to provide reference ideas for intelligent robotic system design.Study 1 used eye-movement techniques to examine the multi-object tracking patterns and eye-movement characteristics of university students without the influence of distractors.In the experiment,participants were required to track 1,3,and 6 moving targets for about 6 s.After the movement stopped,participants were asked to use the mouse to click on the computer screen and report the spatial locations of all targets(i.e.,the full reporting method was used).The results showed that participants exhibited smaller localization errors when the number of targets was less than or equal to four,while the localization error significantly increased when reporting the fifth and sixth targets.Eye movement data showed that participants’ fixation was close to the target object most of the time during tracking,indicating that the participants’ tracking strategy was mainly the target-switching strategy.Study 2 examined the performance of individuals tracking more than four targets simultaneously using a partial report method.In the experiment,participants were asked to determine whether the object that appeared on the screen after the target had stopped moving was the final position of one of the targets that had appeared previously.These positions had a 50% probability of being the previously appearing target position and another 50% probability of not being the target position.The results showed that although the correct tracking rate decreased as the number of targets and the speed of movement increased,the correct rate(80%)was still significantly higher than the guess level when the number of targets was eight,indicating that participants were able to track eight(or even more)targets in the multi-object tracking variant without distractors.Studies 3A and 3B further examined the effect of the tracking time on individuals’ performance in multi-object tracking with and without distractor items.The results showed that in the absence of distractors,individuals’ tracking performance was not affected by tracking time.However,if distractors were present,the tracking performance became worse with the increase in tracking time.This confirms our hypothesis that if a multi-object tracking algorithm with interfering targets is embedded in an intelligent robot,it will increase the computational load on the robot and make tracking less effective.In summary,this study examined the multi-object tracking pattern for college students using a multi-object tracking variant without distractors.Compared to the paradigm with distractors,individuals could successfully track more objects(at least 8)and the tracking was not affected by the length of tracking time.This suggests that the results of the classical tracking paradigm underestimate humans’ multi-object tracking ability due to the large number of cognitive resources consumed by the presence of distractors.Therefore,the corresponding findings may not be used to guide and optimise the design of multi-object tracking algorithms for intelligent robots.On the contrary,multi-object tracking without distractors would be more realistic and the tracking ability of individuals is greatly improved.The corresponding findings have implications for the design of intelligent robots.