This paper discusses dynamics and differences of business models in the car-sharing industry and focuses on open innovation as the trigger of diverse business models among Uber in the U.S., DiDi Chuxing in China, and KakaoT in Korea. We seek to answer the following two questions: What creates the differences in the business models of the car-sharing industry? Do the differences in open innovation motivate the diversity of business models among Uber, DiDi Chuxing, and KakaoT? We incorporated participatory observation, interviews, and semi-structured questionnaire methods in our study. We used two-step participatory observation and interview methods, hence carrying out observation and interviews two times by different researchers with Uber drivers in the U.S., DiDi-Chuxing drivers in Beijing, and KakaoT taxi drivers in Korea to confirm the interview and participatory observation results. First, business models of the car-sharing firms Uber, DiDi-Chuxing, and KakaoT are not fixed but rather are dynamically changing. Second, business models of car-sharing firms are the result of interaction with government regulations, the taxi industry, public transportation, and the automotive car industry. Third, open innovation strategies of car-sharing firms determine the contents and dynamics of car-sharing business models, such as the revenue business model, responsibility business model, and system business model upon interaction with four agencies.

The paper provides accuracy improvement of asynchronous UWB position location algorithm using multiple pulse transmission. We eliminate inter symbol interference by increasing time duration between consecutive pulses, and to reduce AWGN noise we transmitted sufficient number of pulses. At the receiver side we averaged signal over time, and performed correlation with template signal. The result shows that the accuracy of asynchronous position location algorithm improved much compare with its initial signal transmission scheme.

Traditional pedestrian collision warning systems sometimes raise alarms even when there is no danger (e.g., when all pedestrians are walking on the sidewalk). These false alarms can make it difficult for drivers to concentrate on their driving. In this paper, we propose a novel framework for an end-to-end pedestrian collision warning system based on a convolutional neural network. Semantic segmentation information is used to train the convolutional neural network and two loss functions, such as cross entropy and Euclidean losses, are minimized. Finally, we demonstrate the effectiveness of our method in reducing false alarms and increasing warning accuracy compared to a traditional histogram of oriented gradients (HoG)-based system.