移动机器人运动仿真：从Mobotsim到Robotsim

Mobotsim（Mobile Robot Simulator）是早期一款移动机器人仿真软件，在Windows 下的2D动态仿真。它提供了一个图形界面，可以很方便的建立和编辑你的机器人，支持basic语言。
该软件自带一个Basic语言编辑器，所有的控制程序以Basic宏（Macros）的形式提供。
该软件主要用于移动机器人的设计、测试和仿真，可以应用的领域涵盖自动导航技术，避障, 人工智能, a-life, 集成数据传感器等。

---

其优势包括：

• Unlimited number of mobots and obstacles.
• Several obstacle shapes (line, rectangle, round rectangle, arc, ellipse, sector, chord) and free-hand drawing.
• Mobots are differential drive.
• Simulation of Ranging Sensors (tipically ultrasonic sensor)
• Flexible Configuration of Mobots - Platform diameter, wheels diameter, distance between wheels, number of ranging sensors, angle between sensors.
• Configuration of ranging sensors - radiation cone, range, misreading percentage.
• Each mobot has a configurable grid to map the environment. Automatic Certainty Grid method available with use of Ranging Sensor functions.
• Development of simulation through quick and easy BASIC macro writing. Fully Visual Basic for Applications - compatible
• Control program flow and set breakpoints, step-over, step-into, and step-out.
• View debug info such as watch variables, view the call-stack and loaded modules.
• Easy integration of third parties ActiveX controls and dll files. You can add specific tools to the BASIC editor developed to make use of Fuzzy Logic, Genetic Algorithms, Neural Networks, etc.

该软件的局限性也很明显：

• Kinematics, but not dynamics, of the differential-drive mechanism is simulated. Motor errors and wheel slippage are not simulated. Ideal dead-reckoning was assumed. Radiation of Ranging Sensors is approximated by a cone. Angle of incidence and surface features of objects are not considered.

• Despite MOBOTSIM does not provide these hardware simulation capabilities you can make use of the flexibility of BASIC programming to consider these aspects in your simulation, for example, by adding errors intentionally to some variables.

  ---

程序安装后提供了3个典型的仿真例子：

• Wall Follower Robot
• Neural Network Trained Robot
• Virtual Force Field （VFF） Method based Obstacle Avoidance for Mobile Robots
  

  ---

其中，VFF方法的路径规划应用较多，Mobotsim也可以应用于多机器人的路径规划与避障，目前发展出多种算法，包括Traditional VFF based method (T-VFF), Area ratio parameter based method (A-VFF) （参见【1】）, virtual target based method (V-VFF) 等。

仿真示例如下。

单个机器人的路径规划：

两个机器人的路径规划与避碰：

---

鉴于Mobotsim的局限性，以及开发方面的不开放性，近年来出现了多款更贴近于实践的移动机器人运行仿真软件。Mobilesim结合ARIA不失为一个好的手段，同样支持多机器人仿真，但是机器人模型仅能结合公司提供的几种陆地机器人，使用不够灵活。Robotsim是国外某高校研究机构开发的一款开源仿真软件，在一定程度上实现了物理仿真，可定制程度较高。

---

【1】Ni J, Wu W, Shen J, et al. An improved VFF approach for robot path planning in unknown and dynamic environments[J]. Mathematical Problems in Engineering, 2014.