Search Results

Ford Motor Company’s assembly plants build vehicles in a certain sequence. The planned sequence for the plant’s trim and final assembly area is developed centrally and is sent to the plant several days in advance. In this work we present the study of two cases where the plant changes the planned sequence to cope with production constraints. In one case, a plant pulls ahead two-tone orders that require two passes through the paint shop. This is further complicated by presence in the body shop area of a unidirectional rotating tool that allows efficient build of a sequence “A-B-C” but heavily penalizes a sequence “C-B-A”. The plant changes the original planned sequence in the body shop area to the one that satisfies both pull-ahead and rotating tool requirements. In the other case, a plant runs on lean inventories. Material consumption is tightly controlled down to the hour to match with planned material deliveries.

While operating a vehicle in either autonomous or occupant piloted mode, an array of sensors can be used to guide the vehicle including stereo cameras. The state-of-the-art distance map estimation algorithms, e.g. stereo matching, usually detect corresponding features in stereo images, and estimate disparities to compute the distance map in a scene. However, depending on the image size, content and quality, the feature extraction process can become inaccurate, unstable and slow. In contrast, we employ deep convolutional neural networks, and propose two architectures to estimate distance maps from stereo images. The first architecture is a simple and generic network that identifies which features to extract, and how to combine them in a multi-resolution framework. The second architecture is a more specialized one that extracts local similarity information from two images, which are used for stereo feature matching, and fuses them at multiple resolutions to generate the distance map.