摘要
鉴于国内外智能车比赛采用的任务驱动的评价方法,导致参赛队伍为完成任务采取保守的方式,因而背离了技术发展的目的,提出了一种可拓层次分析法与模糊综合评价法相结合的无人驾驶车辆智能行为的模糊-可拓层次分析(Fussy-EAHP)评价法,克服了层次分析法在解决专家经验判断的模糊性问题和判断矩阵一致性方面的不足,有效避免了层次分析法中的大量试算工作,能合理地计算出无人驾驶车辆智能行为各个指标的权重。其量化结果能较好地反映实际情况,且可方便地转化成具体的得分数,便于结果的直观比较或排序。对FC 2011参赛车辆进行评价结果表明,该方法实用、可靠、方便、合理,不仅能对单个指标进行评价,还能对各级指标做出综合评价,并进一步找出无人驾驶车辆各项指标的不足,指明改进的方向。
In view of that both domestic and foreign intelligent vehicle competition adopts task driven evaluation method for intelligent vehicle,leading to a conservative attitude of participant teams for fulfilling task,and hence deviating from the initial objective of technical progress,an evaluation scheme,fuzzy-extended analytic hierarchy process （fuzzy-EAHP）,for the intelligent behavior of unmanned vehicles is proposed by combining EAHP with fuzzy comprehensive evaluation method.It overcomes the ambiguity problem of expertise judgment and the defect of judgment matrix inconsistency in AHP,effectively avoiding massive trying calculation and being able to rationally calculate the weight factors for all indicators of unmanned vehicle＇s intelligent behavior.Its quantitative results can better reflect the actual situation,and can handily be converted to specific scores,facilitating the intuitive comparison and sorting of results.The results of applying the scheme to the evaluation of ＂Future Challenge 2011＂competition show that the proposed scheme is pragmatic,reliable,convenient,reasonable,and can do both the evaluation on individual indicator and the overall evaluation on all indicators,as well as further find out the inadequacies of unmanned vehicle＇s indicators and point out the directions of improvements.
出处
《汽车工程》
EI
CSCD
北大核心
2014年第1期22-27,共6页
Automotive Engineering
基金
国家自然科学基金(90920304)资助
