Based on a biorobotic approach developed in our laboratory over the past 25 years, we have designed and built several terrestrial and aerial vehicles controlling their position and speed on the basis of optic flow cues. In particular, in our project on the autonomous guidance of Micro-Air Vehicles (MAVs) in confined indoor and outdoor environments, we have developed a vision-based autopilot, which is called LORA III (Lateral Optic flow Regulation Autopilot, Mark III). This autopilot, based on the dual optic flow regulation, allows an air vehicle to travel along a corridor by automatically controlling both its speed and its clearance from the walls. The optic flow regulation is a feedback control based on an optic flow sensor, which strives to maintain a perceived optic flow at a constant set-point by adjusting a thrust. The LORA III autopilot consists of a dual optic flow regulator in which each regulator has its own optic flow set-point and controls the robot's translation in one degree of freedom: a bilateral optic flow regulator controls the robot's forward speed, while a unilateral optic flow regulator controls the side thrust, making the robot avoid the walls of the corridor. This autopilot draws on former studies which aimed to understand how a honeybee might be able to center along a corridor, to follow a single wall, and to adjust its speed according to the corridor width. Computer-simulated experiments have shown that a miniature hovercraft equipped with the LORA III autopilot can navigate along a straight or tapered corridor at a relatively high speed (up to 1m/s). The minimalistic visual system (comprised of only four pixels) may suffice for the hovercraft to be able to control both its clearance from the walls and its forward speed jointly, without ever measuring speed or distance, in a similar manner to what honeybees are thought to be capable of. The LORA robot is equipped with two rear thrusters and two lateral thrusters, in addition to the lift fan used to inflate the skirt. The hovercraft can move freely without any umbilicus, which makes its system identification and its own locomotion easier. However, the dynamics of all five motors turned out to be highly sensitive to the drop in supply voltage of the onboard Lithium Polymer (Li-Po) batteries. This is a critical issue for the identification of the robot's dynamical parameters. To perform an efficient system identification of the hovercraft's dynamics, we decided to confer upon each motor a dedicated controller that would make the four thrusters and the lift fan robust to any variations in the battery supply voltage.