本文为葡萄牙里斯本技术大学（作者：Andr′e Guilherme Nogueira Coelho dos Santos）的硕士论文，共96页。

本文研究了智能手机嵌入式系统作为移动机器人导航控制单元的可行性。这项工作可能有助于嵌入式系统领域的进步，通过开发Java嵌入式解决方案，在执行复杂的移动机器人控制导航算法时显示当今的移动电话容量。在这项工作中，我们使用了一个由移动机器人组件（乐高NXT Mindstorms）和由智能手机（诺基亚N80和N95）执行的控制组件组成的系统。智能手机是执行导航算法的嵌入式系统，这些算法的实现考虑了已知的移动限制，例如内存容量和处理速度。实现路径规划、定位和映射的算法分别是：势场、粒子滤波和视觉地标识别。所有的实现都是使用Java编程语言移动版（J2ME）开发的。实验结果表明，尽管所有的硬件限制和算法的移动适应能力都很强，但仍然可以实时执行所提出的算法，并取得满意的效果。该系统能够识别地标，检索可能出现地标的位置，并成功规划其运动避开障碍物。我们证明，当前一代高端手机已经准备好执行如此复杂和苛刻的算法，因此是移动机器人控制的一个可行的解决方案。以目前的容量，这些智能手机可以承担对应的角色，成为我们生活中更重要的一部分。

This thesis studies the feasibility of asmartphone embedded system as a navigation control unit for mobile robots. Thiswork may contribute to the advance of the Embedded Systems field by thedevelopment of Java embedded solutions that show today’s mobile phone capacitywhen executing complex navigation algorithms for mobile robot control. In thiswork we used a system composed by a mobile robotic component (Lego’s NXTMindstorms) and a control component executed by smartphones (Nokia N80 andN95). The smartphone is the embedded system that executes navigation algorithms,which were implemented considering known mobile restrictions and limitations,e.g., memory capacity and processing speed. The algorithms implemented for PathPlanning, Localization and Mapping were: Potential Fields, Particle Filters andVisual Landmark Recognition, respectively. All implementations were developedusing the Java programming language mobile edition (J2ME). The experimentalresults conducted, verified that, although all the hardware restrictions andthe algorithms’s mobile adaptation, it was still possible to execute thereferred algorithms in real-time and with acceptable results. The system wasable to identify landmarks, retrieve possible locations of where it could bepresent and successfully plan its motion avoiding obstacles. We demonstratethat the current generation of high-end mobile phones are ready to execute suchcomplex and demanding algorithms and therefore be a feasible solution formobile robot control. With the current capacity, these smartphones can assumeother roles and become a more essential part of our lives.

1.  引言

2. 移动机器人导航

3. 基于视觉地标识别的地图绘制

4. 基于粒子滤波的定位

5. 基于势场的路径规划

6. 嵌入式系统原型

7. 实验结果

8. 结论

附录A 潜在字段XML输入

附录B 原型硬件的规格参数

更多精彩文章请关注公众号：