Autonomous vehicles

Fuel cells have attracted considerable interest because of their potential for efficient conversion of the energy (AG) from a chemical reaction to electrical energy (AE). This efficiency is achieved by directly converting chemical energy to electricity. Conventional systems burn fuel in an engine and convert the resulting mechanical output to electrical power. Potential applications include stationary multi-megawatt power plants, battery replacements for personal electronics, and even fuel-cell-powered unmanned autonomous vehicles (UAVs). 

The SeaDog is also available in two configurations. A handheld version can be easily used by divers, while the AUV model is specifically configured for integration with underwater autonomous vehicles, such as the REMUS (Remote Environmental Monitoring Unit) shown in Figure 9.8. 

Broggi, A., Bertozzi, M., Fascioli, A., Conte, G., Automatic Vehicle Guidance The Experience of the ARGO Autonomous Vehicle, World Scientific, Singapore, 1999. 

Future prospects for research on deep-sea hydrothermal vents and cold seeps are very promising. New technical approaches and exploration methods (e.g. autonomous vehicles like ABE Embley etal., 2002) will likely benefit this research. Molecular methods will help us to better understand the diversity and possibly the metabolisms of un-culturable environmental microorganisms. Long-term seafloor observatories like NeMO (Embley Baker, 1999) and NEPTUNE (Delaney etal., 2000) will allow us to more easily study episodic events like volcanic eruptions at ridges and massive methane releases from seeps. As well, these observatories will provide data in the form of long-term time... 

Figure 7. Detection of a TNT plume in the marine environment with the sensor mounted on an underwater autonomous vehicle.

There are also other types of automatic vehicles that are called robots - shuttle trains, experimental terrain vehicles, and underwater autonomous vehicles. Although they are very sophisticated, these machines depart to far from the notion of a robot being humanlike in its appearance and/or behavior. So in order to not make the term robot synonymous with the terms automatic or autonomous, one should avoid including robot in the designation of these machines. 

Bond graph-based quantitative FDI methods use ARRs [45 8], parameter estimation in the case of multiple simultaneous faults [46, 49], or observers [34]. Recently, bond graph-based FDI has been applied to various systems such as an industrial steam generator [50], an industrial chemical reactor [51], a mobile robot test bed [52], or an intelligent autonomous vehicle [53]. 

In any case, FDI is a prerequisite also for this task and bond graph based ARR residual generation can provide the information needed by fault diagnosis. Chapter 11 of reference [11] addresses fault tolerant control of systems represented by continuous time model and related issues such as system inversion. In [12], a bond graph approach to diagnosis and FTC has been recently presented and applied to an intelligent autonomous vehicle. FTC of hybrid systems has been considered for instance... 

Loureiro, R. (2012). Bond graph model based on structural diagnosability and recoverability antilysis Application to intelligent autonomous vehicles [PhD thesis]. L Universite Lille 1. 

W. Nelson, 1. Cox, Local path control for an autonomous vehicle, in Proceedings of the IEEE... 

Risk contours may also be a useful concept for planning safe trajectories for fully autonomous vehicles and not just collision mitigation strategies. In future where there will be different levels of automation, risk contours may be one way to assert that the transitioning process from self-driving mode to automated model is safe because it has a notion of safe distance in form of timing. 

FIGURE 3 Left Map of the DUC course (lines = map circles = stop signs). Black squares indicate where brakes were applied quickly during the six-hour mission. Right A model helicopter (operated by remote control or as an autonomous vehicle) in mid-maneuver. These complex maneuvers can be learned from an expert or by experimentation. Photo by E. Fratkin. 

How, J., B. Bethke, A. Frank, D. Dale, and J. Vian. 2008. Real-time indoor autonomous vehicle test environment IEEE Control Systems Magazine 28(2) 51-64. 

Pongpunwattana A., Rysdyk R, Evolution-based Dynamic Path Planning for Autonomous Vehicles, Innovations in Intelligent Machines, pp. 113-145, 2007... 

There is much work afoot to bring autonomous robots (AR) into public spaces, both on the ground and in the air. We therefore need high confidence that their behaviour will be safe. This is difficult, however, given the complexity of environments that the robots must interact with, and the scope of authority that they need in order to effectively respond to those environments. There have already been minor accidents caused by the control software of autonomous vehicles, for example in the DARPA Urban Challenge [1], and there are likely to be more as their numbers increase. We need ways to verify and validate such control software, finding ways in which it could cause an accident, whether due to intrinsic algorithmic limitations or implementation-specific faults. 

Liunelsky, V., Stepanov, A. (1984). Effect of imcertainty on continuous path planning for an autonomous vehicle. In Proceedings of the 23rd IEEE Conference on Decision and Control, (Vol. 23, pp. 1616-1621). IEEE Press. 

Autonomous vehicles are typically controlled from a control system that is composed of sensor units, actuators, and a processing center. Sensors and actuators can be connected to a single processing system forming a centralized architecture or related to different processing units interconnected to each other forming a distributed architecture. 

Automated Electric Monorail. In this system, the carriers are not linked in a continuous chain, but are autonomous vehicles. They can also operate in an inverted position, carrying loads through curves, switches, and different elevations. The system is especially suited for applications where carrier and inventory identification is critical, and precision positioning or robot interface is required. By the use of individual carriers, the load in each carrier can be tracked and controlled at all times. 

---