2.1. PROBLEM FORMULATION 9
Table 2.1: Driving style recognition results in suv cycles
Aggressive Cycle Moderate Cycle Conservative
Cycle
Acceleration 0.55 (149) 0.43 (113) 0.34 (106)
Brake 0.58 (33) 0.56 (25) 0.36 (22)
Cruise 0.83 (149) 0.69 (126) 0.70 (124)
Turn 0.41 (6) 0.29 (7) 0.29 (7)
Agressive Profile Moderate Profile Conservative Profile
Ref Acc (m/s
2
)
Ref Acc (m/s
2
)
Ref Acc (m/s
2
)
Ref Spd (km/h)
Veh Spd (
km/h)
Veh Spd (
km/h)
Veh Spd (
km/h)
R
ef Spd (
k
m
/h)
R
ef Spd (
k
m
/h)
Figure 2.4: Pre-defined 3D reference acceleration profiles.
Finally, the consistency and robustness of the algorithm are verified using the test data set. e
test shows consistency in the identification and aligns with drivers’ perception. e above test-
ing results validate the suitability of this approach for DSR, its onboard implement capability
and robustness to vehicle and driver characteristics. More detailed algorithms with experimental
results can be found in [34].
Based on the above recognition and classification algorithms, the features of aggressive,
conservative, and moderate driving styles can be extracted, and online recognition of a driver’s
driving style can be realized using the well-trained model as well. Meanwhile, according to the
above features obtained, the three dimensional human-like acceleration profiles are developed
for each driving style, as illustrated in Fig. 2.4.
2.1.5 REQUIREMENTS FOR THE DESIGN AND OPTIMIZATION OF
CPVS
e requirements for vehicle design and control involve dynamical performance, energy effi-
ciency, and ride comfort. Driving style consideration implies the introduction of multiple trade-
offs between performances that are set as the objective functions in our optimization problem,
under different driving styles, operating conditions, and driving tasks.
..................Content has been hidden....................
You can't read the all page of ebook, please click here login for view all page.