Over the past few years, mobile robots have started emerging as one of the most important assets in industries for material handling and other intra-logistics operations. These robots have minimized the need for manual handling and an increased number of handling tasks. Little do people realize that even though these robots carry heavy loads but still face functional challenges that are not often noticed and have adverse effects on their maneuverability.
Challenges Synoptic
Mechanical Design Overview
The design of mobile robots capable of intelligent motion and action involves the integration of many different bodies of knowledge. The aim of this system is to idealize an existing autonomous mobile robot, on all levels. This includes the mechanics, kinematics, dynamics, perception, sensor fusion, localization, path planning, and navigation. All these aspects must be reviewed and modified to a modular system if necessary new modular modules must be designed and developed. This way a robust and modular autonomous mobile robot, capable of intelligent motion and performing different tasks will arise.
The major challenges include mechanical structure, navigation, and human-centred intelligent control, of which navigation is the most challenging functionality required for such autonomous systems. The navigation comprises four dominant blocks of competencies: perception, localization, cognition and motion control.
- Perception is the ability of a robot to interpret meaningful data from its sensor,
- Localization defines how good a robot determines its position in the environment,
- Likewise, Cognition and Motion control helps in extracting a way to achieve its goal and modulating the motor controller to reach the desired trajectory.
Of all the above four competencies in navigation, localization is considered the most challenging area which requires the greatest research attention.
A Pitfall to Mobile Robots
Assuming one could just attach a GPS sensor to a mobile robot that could solve the localization problem informing the robot of its exact position in the environment. Unfortunately, the current GPS system is not practical with accuracy to say several meters which are almost unacceptable for localizing the mobile robot. Furthermore, the current advancements with positioning technologies are not proving any place in the market especially when it comes to indoors or in obstructed areas. Also, localization is not just limited to determining an absolute pose in space, rather a series of collaborative tasks like building a map, then identify the robot’s relative pose with respect to mapping. In other words, one can say that the robot’s sensors play a crucial role in the localization and the sensor’s inaccuracy and incompleteness contributes to major challenges in localization.
In perception, the major contributor lies with sensor noise and aliasing, further aggravating the problem of localization. On the other hand, using a noise-free sensor alone can’t solve these challenges of insufficient information to identify robot’s pose in the world, instead, it also requires robot programming to recover the robot’s position over time based on series of sensor readings.
