Predictive virtual lane method using relative motions between a vehicle and lanes

Abstract

We propose a new approach for virtual lane prediction. The main contribution of the proposed method is that the predicted virtual lane can be substituted for lane detection using a camera sensor when the camera image processing fails to detect the lane. The proposed method generates the predicted virtual lane using the relative movement between a vehicle and a lane. To predict the lane, a third-order polynomial function of the longitudinal distance is used as a lane model. Each coefficient of the lane polynomial function at the next sampling time is geometrically calculated using the relative movement of a vehicle, the lanes, the longitudinal velocity and the yaw of the vehicle at the present time. Then, the predictive virtual lane at the next sampling time is obtained without the lane information from the camera sensor at the next sampling time. The proposed method is simple enough that it is suitable for real implementation. The performance of the proposed method was evaluated via experiments with a test vehicle.

A comparative study of lane keeping system: Dynamic and kinematic models with look-ahead distance

“…Under the operation condition of LKS or LXC, the effect of the lateral acceleration of vehicle at c.g would be small enough to be neglected (a y (k) ≈ 0) [18], [19]. Then from (7), one can obtain the following discrete time state space representation using the state vector x(k), control input u = δ f and external signals…”

Lane estimation using a vehicle kinematic lateral motion model under clothoidal road constraints

“…Clothoid sprials are curves that are used for smooth transition from one curvature value to another. The curvature of clothoidal curve at arc length, s, is described as follows κ = 6c 3 s + 2c 2 (12) where 2c 2 is the road curvature at s = 0 and 6c 3 is its rate. Integration of this equation twice with respect to s leads to a clothidal cubic polynomial curve road model:…”

“…Parameters used in the simulation study are nominal values of a test vehicle. Following three cases are simulated and all cases use LKS module which is already developed in [12], [13] Fig 5, when the camera returned to normal condition, the Fail method tends to exert large steering wheel angle to keep the lane. However, the Comp method did not have such a large steering wheel angle and this result in more comfortable ride to passengers rather than the Fail method.…”

“…Another approach is generating predictive virtual lane (PVL) using data from the inertia measurement unit and relative kinematic motion under the assumption that the lane is fixed. It is reported that PVL can cope with up to several sampling period camera failures providing improved lateral control performance of lane keeping system (LKS) [12], [13]. Weather condition, lighting changes and complex shadowing often happen in many situations and these are critical problems of vision processing [8], [14].…”

“…The kinematics model was reformulated with a look-ahead distance to design a multirate estimator. The clothoidal road constraints are used to resolve a longer failure than that considered in the previous study [12], [13]. Furthermore, we design a decentralized multirate estimator to resolve aforementioned irregular and slow update period of the vision sensor.…”