Search Results

An efficient approach to classify time series physical measurement data of shift control of automatic transmission for vehicles is presented. Comfortable acceleration is the essential factor of today’s vehicles. Shift control of automatic transmission of vehicles directly contributes to the comfortable acceleration. Since calibration of automatic transmission of vehicles is time consuming task for expert engineers, the development of autonomous calibration is desired to reduce product development period in today’s competitive automobile market. In the stage of product development, it is difficult to obtain a large amount of physical measurement data. Therefore, we need to develop machine learning method for limited amount of data. For this purpose, we develop the method to classify time series measurement data of shift control of automatic transmission of vehicles. We use support vector machine (SVM) as a machine learning technique.

In recent years, the development time of vehicles has further accelerated, and automation of the development is an urgent task. One example of time wasting tasks is gear-shift calibration. For this purpose, Kawakami et al. have studied OK/NG classification of shift quality by using neural networks. However, their classifiers have a problem in versatility over different AT hardwares. In this paper, we develop autoencoders to realize similar/not-similar classification on three AT hardwares of vehicles. These hardwares have different lock-up multi/single-plate clutch structures. Experimental results show that the performance of similar/not-similar classification is high in terms of AUC.

A newly developed classifying method for time series measurement data of automatic transmission of vehicles is presented. The proposed method uses deep convolutional neural networks to learn what is comfortable acceleration. In addition, our proposal method employs supervised learning based on experienced engineer’s criterion. As a demonstrative problem, we consider the classification of time series measurement data for lock-up clutch control of an 8 speed automatic transmission.

This paper presents a study on a design methodology for lightweight automatic transmission parts for vehicles. To improve fuel economy of the vehicle, weight reduction is a primary issue when designing automobile parts. Topology optimization, a particularly flexible type of structural optimization method, is an effective tool for developing lightweight designs. However, topology optimization methods available in most commercial software often provide optimization results that include grayscale areas, which cause the designs to be impractical from an engineering and manufacturing standpoint. To overcome this difficulty, we apply a level set-based topology optimization method that yields clear optimal configurations and therefore facilitates further improvement of the product design by engineers. In addition, we consider the imposition of a uniform cross-section constraint to improve the manufacturability of the obtained optimization results.

Efficient method to solve large-scale eigenvalue problem in vibration is presented. NVH (Noise, Vibration and Harshness) performance is an important quality measure of vehicles. Therefore, the reduction of vibration is one of the key considerations of new automatic transmission design. In addition, reduction of product design time is another important requirement. Computer Aided Engineering (CAE) is becoming a more important methodology to reduce product design time. However, computational time of eigenvalue problem takes long. We propose parallel eigenvalue computation (Sakurai-Sugiura method) for large-scale eigenvalue problems. This method has a good parallel scalability according to a hierarchical structure of the method. As the demonstrative problem, we consider large-scale computation of eigenvalue problem for AISIN AW FWD automatic transmission.

A design process to improve the structural performance under contact conditions of automatic transmission of vehicles is presented. The proposed process uses newly developed nonlinear contact optimization techniques. Additionally, useful structural optimization techniques such as 3D topography and the beta method are described. As a demonstrative problem, we consider the improvement of a 6- speed AISIN FWD automatic transmission which is subjected to static loadings and nonlinear contact conditions.

A design process to reduce mass and sound pressure for automatic transmissions of vehicles is presented. The proposed process uses a newly developed topography optimization technique. Additionally, useful and necessary techniques including sound pressure optimization and the beta method are described. As a demonstrative problem, the design process is successfully applied to reduce mass and sound pressure of a 6 speed AISIN AW FWD automatic transmission.

A methodology to optimize sound pressure responses of a structure, producing a radiation noise due to structural vibration, is presented. The method involves a finite element analysis module to calculate structural vibration, an optimization module to perform sensitivity analysis and structural optimization, and an acoustic module to compute acoustic transfer vectors. The proposed design system is successfully implemented and is demonstrated in the paper using several example problems.

A design process to improve noise and vibration (NV) performance for automatic transmissions of vehicles is presented. The proposed process uses a newly developed large-scale optimization technique to solve large-scale analysis problems. Additionally, useful and/or necessary techniques including Topometry optimization, beta method and modal contributions are described and used to solve modal dynamic optimizations. As a demonstrative problem, the process is successfully applied to improve the NV performance of the AISIN AW FWD automatic transmission.