English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Tools annual

Global: Zendy Free Plan

Global: Zendy Tools monthly

Global: Zendy Plus monthly

Autonomous car simulating is both safer and more economical than directly using real car for self-driving related research work. In this paper, we propose a framework that fully utilize the Lego Ev3’s sensor and Android App to simulate an autonomous car. First, the App receives sensor data from Lego Ev3 and images from Android phone, then using specified strategies to guide computing these data. Finally, when the computation is completed, App makes decisions to order Lego Ev3 avoiding obstacles in front of it. The computing process contains image process, simultaneous mapping construction and multi data fusing, which is highly related to autonomous car technology, but safer and more economical. Experimental results show that the proposed framework can be used for autonomous car simulating, so we can do more autonomous car related research works on this framework in the future. Introduction Autonomous car or self-driving car has become a hot research pot in recent years. However, it is so expensive to buy all the equipment for autonomous car driving and testing. In addition, it is unsafe to test an autonomous car in an open environment. Fortunately, the algorithms behind the autonomous car driving are the same, so we can use some other equipment to simulate the self-driving environment and equipment. Moreover, we have several Lego Ev3s in our laboratory, so we can use Lego to simulate an autonomous car. Lego Ev3 provides several sensors, such as Ultra Sound Sensor, Gyro Sensor, Touch Sensor and Color Sensor, so it is very convenient for us doing research work. Unfortunately, due to the limitation of its computational efficiency, it cannot be used to deal with any images. Usually, image processing is a fundamental part of an autonomous car, thus we cannot directly use Lego Ev3 for simulating the autonomous car. With the fast progress achieved in IT, especially in mobile computation environment, we can use mobile phone to deal with larger data than before. In addition, the camera is the basic component of a high-end mobile phone, and Android becomes the most popular OS in mobile phone field. So rather than simply using Lego Ev3 as the simulator, we can now use Android phone together with Lego Ev3 to simulating an autonomous car. The key issues by using the Lego Ev3 and Android phone for simulating an autonomous car are as follows. First, it is difficult for android phone directly communicating with Lego Ev3, because the operating system on Lego Ev3 is a closed one. Second, the mono camera on the Android phone cannot easily provide depth information that will be used for obstacle avoidance. Therefore, there are so many previous research works focusing on these two issues. Abdullah et al. [1] proposed an autonomous vehicle convoy algorithm. The proposed algorithm utilized color sensor and ultra sound distance sensor to detect the road and the distance between the front cars. Unlike the real autonomous car, it can only follow a predefined route, so it is insufficiently simulating an autonomous car. 495 Radcliffe et al.[2] recently proposed a machine vision algorithm for orchard navigation. In their paper, Lego Ev3 is performed as the controller, and a laptop computer is used for image processing. Since the laptop is energy consuming, it could not work for a long time. In such a circumstance, Android phone is better than laptop. Chen et al.[3] proposed a graphical based programing language named VIPLE to control different robots, such as Lego Ev3, Raspberry PI, Intel Edison etc.. VIPLE is a well-developed IDE, so users can easily manipulate it. However, it is impossible to deal with the image in the latest version. In this paper, we propose a new framework that can be used for autonomous car simulating. It contains three main parts. The first part deals with the connection problem between Lego Ev3 and the Android phone, the second part processes the information gathered by the Android phone camera. The last part fully utilizes the information gathered from Android phone and Lego Ev3 to guide Lego Ev3 as an autonomous car. The rest of the paper is organized as follows. Section 2 gives the detail information of the proposed simulating framework. Section 3 analyses the experimental results. The last section concludes all. Autonomous Car Simulating Based on Lego Ev3 and Android App Autonomous car must have three fundamental functions, sensor data inputting, information processing and action. Therefore, when we are trying to use Lego Ev3 and Android phone to simulate an autonomous car, the urgent problem we faced is that how to realize the three parts that are equivalent to the real autonomous cars’. In this section, we first talk about the information acquiring process, and then make some introductions on information processing, especially on image processing. Finally, we concern on the reaction of the Lego Ev3. The framework is shown in Figure 1. Figure 1. Framework of the simulating system. Sensor Data Inputting We can use Ultra sound sensor to get the distance information between Lego Ev3 and the obstacle and turning angles from Gyro sensor. Meanwhile, we receive images from Android phone and try to use OpenCV[4] or other SDK to process these images in an Android App. In summary, there are three types of data will input into our framework. Before we are going to talk the details of the inputting step, we will firstly introduce the proposed framework. In our framework, Lego Ev3 just plays a role as the actor and information provider, because we do not directly use any computation resource of Lego Ev3. The core of the framework is an APP that running on the Android phone. It connects to Lego Ev3 through Bluetooth and communicates with laptop through WIFI. The APP deals with all the sensor data including the images from Android phone, and then gives the orders that Lego Ev3 should turn into action. Android APP establishes a connection between Android phone and Lego Ev3 through Bluetooth. It is not an easy work for APP to communicate with Lego Ev3, because Lego Ev3 did not provide any open SDK or APIs. The only thing we can do is that trying to follow the command’s format specified by LEGO to build a command string, and then sending the string to Lego Ev3 through Bluetooth connection. Although sending a command string is an easy task, the execution result is unknown. In