Page 447 - Automotive Engineering Powertrain Chassis System and Vehicle Body

CHAP TER 1 4. 2 Decisional architecture

above-mentioned problem and to increase the robustness sensors are emitting/receiving signals at each time step
of the ‘target tracking’ behaviour. The chosen approach for sensing each side of the car). The minimal number of
mainly consists in coupling the controller with an on-line ultrasonic sensors required by the parallel parking SBM is
‘local trajectory generator’ which tries to continuously eight: three for looking in the forward direction, two
evaluate the sequences of states – i.e. the (position, ori- located on each side of the vehicle and one for looking in
entation, velocity) parameters – of the leading vehicle on the backward direction. This ultrasonic sensor system is
a short time interval (instead of only using an in- intended to test the control algorithms for low-speed
stantaneous approach). This approach allows us to still motion only; a more complex sensor system (e.g. a com-
control the motions of the following vehicle, when the bination of vision and ultrasonic sensors) should be used
target has been lost for a short time period. Current work to ensure reliable operation in a more dynamic environ-
deals with the processing of exceptions and of the car ment. This is the purpose of current work.
entrance and exit procedures. The CCD-camera has a resolution of 2048 pixels, and
it operates at a rate of 1000 Hz; it is equipped with
14.2.4 Experimental results a cylindrical lens and an infrared polarized filter. This
device operates in relation with an infrared target for
localizing the leading vehicle in the platooning SBM; this
14.2.4.1 Experimental vehicles
target is made of three pulsing sets of LED organized
along vertical lines, as shown in Fig. 14.2-22.
The approach described in this chapter has been imple- The motion controller of our control and decisional
mented and tested on our experimental automatic architecture monitors the current steering angle, loco-
vehicles (a modified Ligier electric car, and the Cycab motion velocity, travelled distance, coordinates of the
electric vehicle designed and developed at INRIA (Baille vehicle and range data from the environment, calculates
et al., 1999). These vehicles are equipped with the main appropriate local trajectories and issues the required
following capabilities:
servo commands. It has been implemented using the
1. a sensor unit to measure relative distances between Orccad software tools (Simon et al. 1993) running on
the vehicle and environmental objects, a workstation; the compiled code is transmitted via
2. a servo unit to control the steering angle and the Ethernet to the control unit operating under the
locomotion velocity, VxWorks real-time operating system.
3. a control unit that processes data from the sensor and
servo units in order to ‘drive’ the vehicle by issuing 11.4.2 Experimental run of the trajectory
appropriate servo commands.
following manoeuvre
These vehicles can either be manually driven, or they can
move autonomously using the control unit; this unit is An experimental run of the trajectory following SBM
based on a Motorola VME162-CPU board and a trans- with obstacle avoidance on a circular road (roundabout)
puter network for the Ligier, and on a distributed control is shown in Fig. 14.2-17. In this experiment, the Ligier
architecture implemented using a CAN bus and micro- vehicle follows a nominal trajectory along the curved
controllers for the Cycab (Baille et al., 1999). The sensor traffic lane, and it finds on its way another vehicle moving
unit of the vehicle makes use of a belt of ultrasonic range at a lower velocity (see Fig. 14.2-17(a). When the moving
sensors (Polaroid 9000) and of a linear CCD-camera. obstacle is detected, a local trajectory for a right lane
The servo unit consists of a steering wheel servo-system, change is generated by the system, and the Ligier per-
a locomotion servo-system for forward and backward forms the lane changing manoeuvre, as illustrated in
motions, and a braking servo-system to slow down and Fig. 14.2-17(b). Afterwards, the Ligier moves along
stop the vehicle. The steering wheel servo-system is a trajectory parallel to its nominal trajectory, and a left
equipped with a direct current motor and an optical lane change is performed as soon as the obstacle has been
encoder to measure the steering angle; the locomotion overtaken (Fig. 14.2-17(c)). Finally the Ligier catches up
servo-system is equipped with an asynchronous motor its nominal trajectory, as illustrated in Fig. 14.2-17(d)).
and two optical encoders located onto the rear wheels The corresponding motion of the vehicle is depicted in
(for odometry data); the Ligier vehicle is also equipped Fig. 14.2-18(a). The steering and velocity controls applied
with a hydraulic braking servo-system. during this manoeuvre are shown in Fig. 14.2-18(b) and
The ultrasonic range sensors used in the described Fig. 14.2-18(c). It can be noticed in this example that the
experiments have a measurement range of 0.5–10.0 m, velocity of the vehicle has increased when moving along
and a sampling rate is 60 ms. The sensors are activated the local ‘parallel’ trajectory (Fig. 14.2-18(c)); this is due
sequentially in order to make more robust measurements to the fact that the vehicle has to satisfy the time con-
in the different regions defined by the vehicle i.e. four straints associated with its nominal trajectory.

454