Gait, the walking pattern of individuals, is one of the most important biometrics modalities. Most of the existing gait recognition methods take silhouettes or articulated body models as the gait features. These methods suffer from degraded recognition performance when handling confounding variables, such as clothing, carrying and view angle. To remedy this issue, we propose a novel AutoEncoder framework to explicitly disentangle pose and appearance features from RGB imagery and the LSTM-based integration of pose features over time produces the gait feature. In addition, we collect a Frontal-View Gait (FVG) dataset to focus on gait recognition from frontal-view walking, which is a challenging problem since it contains minimal gait cues compared to other views. FVG also includes other important variations, e.g., walking speed, carrying, and clothing. With extensive experiments on CASIA-B, USF and FVG datasets, our method demonstrates superior performance to the state of the arts quantitatively, the ability of feature disentanglement qualitatively, and promising computational efficiency.

Choice models play a critical role in enterprise-driven design by providing a link between engineering design attributes and customer preferences. However, existing approaches do not sufficiently capture heterogeneous consumer preferences nor address the needs of complex design artifacts, which typically consist of many subsystems and components. An integrated Bayesian hierarchical choice modeling (IBHCM) approach is developed in this work, which provides an integrated solution procedure and a highly flexible choice modeling approach for complex system design. The hierarchical choice modeling framework utilizes multiple model levels corresponding to the complex system hierarchy to create a link between qualitative attributes considered by consumers when selecting a product and quantitative attributes used for engineering design. To capture heterogeneous and stochastic consumer preferences, the mixed logit choice model is used to predict consumer system-level choices, and the random-effects ordered logit model is used to model consumer evaluations of system and subsystem level design features. In the proposed approach, both systematic and random consumer heterogeneity are explicitly considered, the ability to combine multiple sources of data for model estimation and updating is provided using the Bayesian estimation methodology, and an integrated estimation procedure is introduced to mitigate error propagated throughout the model hierarchy. The new modeling approach is validated using several metrics and validation techniques for behavior models. The benefits of the IBHCM method are demonstrated in the design of an automobile occupant package.

This paper presents an algorithm for generating tolerance chains from the mating relations between components of assemblies. The algorithm is developed upon a feature-based assembly modeling strategy that represents each component in close relation to its mating features, dimensions, and tolerances. The mating relations within an assembly are described by a mating graph. Tolerance chains together with their dimensions and tolerances are generated automatically by searching through a mating graph for matching mating features. A prototype program package based on the presented algorithm has been developed, and several examples of various complexity have been tested with success.

Demand models play a critical role in enterprise-driven design by expressing demand and revenues as a function of price, and product attributes. Revenues and cost, expressed as a function of product attributes, provide the basis for predicting profits; the primary objective of corporate decision-making. However, existing demand modelling approaches in the design literature do not sufficiently address the unique issues that arise when complex systems are being considered. Current approaches typically consider customer preferences for only quantitative product characteristics and do not offer a methodology to incorporate customer preference-data from multiple component/subsystem-specific surveys to make product-level design trade-offs. In this paper, we propose a hierarchical choice modelling approach that addresses the special needs of complex engineering systems. The approach incorporates the use of qualitative attributes and provides a framework for pooling data from multiple sources. Heterogeneity in the market and in customer-preferences is explicitly considered in the choice model to accurately reflect choice behaviour. Ordered logistic regression is introduced to model survey-ratings and is shown to be free of the deficiencies associated with competing techniques, and a Nested Logit-based approach is proposed to estimate a system-level demand model by pooling data from multiple component/subsystemspecific surveys. The design of the automotive vehicle occupant package is used to demonstrate the proposed approach and the impact of both packaging design decisions and customer demographics upon vehicle choice are investigated. The focus of this paper is on demonstrating the demand (choice) modelling aspects of the approach rather than on the vehicle package design.

The issue of moving from a mass production operating mode to mass customization, or even limited customization, has many companies struggling to reorganize their product architectures. Enabling the production of several related products for different market segments, from a common base, is the focus of the product variety design research area. In this paper, the applicability of product variety design concepts to the design of automotive platforms is explored. Many automotive companies are reducing the number of platforms they utilize across their entire range of cars and trucks in an attempt to reduce development times and costs. To what extent can research on product variety design apply to the problem of platform commonization? This question is explored by comparing product variety design concepts (standardization, modularity, mutability, etc.) to platform structures and requirements. After assessing the applicability of these concepts, a platform representation and methods for measuring platform commonality are proposed that incorporate key characteristics of these concepts. An application to two platforms is included. Although preliminary, this work has led to insight as to why automotive platform commonization is difficult and how product design variety research can potentially aid commonization. The findings are potentially applicable to product platforms in general.