A clear standard for acceptable behavior of an autonomous vehicle in the UK road environment touches on three main areas. The first part is the interaction between the user and the automated car. On this area, it is demanded that the vehicle obeys the instructions of the user at all times to avoid unnecessary conflict. The second area is the interaction between the road users such as pedestrians with the automated car. Under this area, the automated car is expected not to injure other road users either through inaction or through their actions. Finally the attitude of the public towards these automated cars. On this area, the autonomous car is expected to behave in a way that its operations do not harm or affect the public thus winning the trust and acceptability of the public. An acceptable attitude of the public is important of the vehicles are to become a success.

Taxonomy of faults refers to the threats that may affect a system. These threats can be categorized into different groups depending on their nature, effect on system, location, and development among other things. A taxonomy of faults that affects the system during its lifetime can be categorized into developmental, operational, internal, external, natural, hardware, software, malicious, non-malicious, deliberate, non-deliberate, accidental, incompetence, transient, and permanent faults (Avizienes, Laprie, & Randell, 2010). In the real world, autonomous vehicle failure is majorly connected to system failures. An example is when the vehicle experienced difficulties in sensor data thus leading to a situation where phantom obstacles and slow moving vehicles could not be identified (Fletcher et al., 2008). Another example is vehicle failure caused by the inability to anticipate vehicle intent (Fletcher et al., 2008). In both cases, the system set-up of the vehicles was prone to faults that were likely to cause disruption and accidents.

To achieve better behavior in the vehicle controller, the controller needs to provide appropriate directives mainly on distance keeping aggressiveness and lateral bias. There is need to build actuation mechanisms and sensors so as to maintain control. For example, an Adaptive Cruise Control (ACC) controller can be installed to control the throttle and brake to guarantee that the car follows the leader. With the brakes and throttle controlled by the ACC, cases of aggressiveness from behind and inability to spot slow moving cars will reduce. This is an important addition because it ensures that the car operates smoothly, is stable, and safe. In addition, the installation of a lane centered module will ensure that the car operates within its lane and is able to send signals to other road users when changing lanes. However, there is need for motion planner algorithms that will ensure that the autonomous vehicles are able to slow down and merge into the next lane.

The current vehicle controller exhibits two main failures. The first failure is system failures. This form of failure originates from development of the system (Alexander et al., 2015). The system fails to address pertinent issues such as the identification of other vehicles and the threat they pose. The second failure is permanent fault where the controller was developed with a fault that cannot be fixed or improved but only be replaced by a new controller. In my newly developed work, detection of failures takes two approaches. The first approach is testing based on the scope of the controller. This kind of testing focuses on the space within which the controller operates and if the environment is conducive for its effective operation (Alexander et al., 2015). The second approach is testing based on the objective of the controller. The question to be answered in this case is if the controller fits the purpose it was intended for.