删除或更新信息,请邮件至freekaoyan#163.com(#换成@)

运行数据驱动的手机性能需求推断与感知分析

本站小编 Free考研考试/2022-02-12

王峥1,褚学宁1,陈汉斯1,张磊1,颜波1,刘航2
1. 上海交通大学 机械与动力工程学院, 上海 200240;2. 郑州航空工业管理学院 管理工程学院, 郑州 450000
出版日期:2018-07-28发布日期:2018-07-28
通讯作者:褚学宁,男,教授、博士生导师,E-mail: xnchu@sjtu.edu.cn.

基金资助:国家自然科学基金资助项目(51475290,51075261)

A Method for Prediction of Smartphone Performance Requirements and Perception Analysis Based on Operating Data

WANG Zheng,CHU Xuening,CHEN Hansi,ZHANG Lei,YAN Bo1,LIU Hang
1. School of Mechanical Engineering, Shanghai Jiao Tong University, Shanghai 200240, China;2. Collage of Management Engineering, Zhengzhou Institute of Aeronautical Industry Management, Zhengzhou 450000, China
Online:2018-07-28Published:2018-07-28







摘要/Abstract


摘要: 针对传统用户需求分析方法存在的主观性强、需求描述难以量化等问题,提出了基于运行数据的性能需求推断方法.首先采用分布拟合法对运行性能数据进行拟合和估计.在此基础上,确定用户需求区间,并以累计概率分布的逼近函数——Sigmoid-like来构建用户需求满足度函数.基于该函数提出了用户性能需求推断公式,并将推断结果与用户感知结果进行对比分析.最后,以手机电池容量性能需求为例,利用自主研发的Smart Monitor系统采集到的手机日耗电量、电池剩余电量、充电次数等数据,对志愿者的手机电池日耗电量进行2次推断实验.2次实验的结果误差均小于 5.00%,验证了所提出的理论方法具有可靠性.
关键词: 性能需求分析, 感知分析, 产品运行数据, 数据驱动, 需求满足度函数
Abstract: Customer requirement analysis is the premise of product design. The traditional methods for requirement analysis include market research methods and online comment analysis. These methods are subjective and difficult to quantify the description of the requirements. Therefore, a method based on operating data is proposed to identify the customer performance requirements. Firstly, the distribution fitting method is used to fit the performance distribution. Based on this, the user required performance interval is determined, and the customer satisfaction function is constructed with the approximation of Sigmoid-like function of the cumulative probability distribution. Based on the function, for a given certain satisfaction degree, the customer required performance can be obtained and also used for comparison with user’s perception. At the end of this paper, two experiments are conducted for analyzing the volunteers’ performance requirement on their mobile battery module through the data collected by the Smart Monitor system we developed, including daily cumulative power consumption, residual capacity, number of charge cycles. The errors between two experimental results are less than 5.00%, which verifies the theoretical method we proposed.
Key words: performance requirement analysis, perception analysis, operating data, data driven, customer satisfaction function


PDF全文下载地址:

点我下载PDF
相关话题/数据 工程学院 实验 上海交通大学 机械