A novel data-driven rollover risk assessment for articulated steering vehicles using RNN

Abstract

Articulated steering vehicles have outstanding capability operating but suffer from frequent rollover accidents due to their complicated structure. It is necessary to accurately detect their rollover risk for drivers to take action in time. Their variable structure and the variable center of mass exhibit nonlinear time-variant behavior and increase the difficulty of dynamic modelling and lateral stability description. This paper proposes a novel data-driven modelling methodology for lateral stability description of articulated steering vehicles. The running data is first collected based on the typical operations that prone to rollover and then classified into two types: Safety and danger. The data quality is further improved by wavelet transformation. Finally, an RNN model is built on the data. The experimental results show that the output of the RNN model can accurately quantify lateral stability of the vehicle, i.e., the risk of rollover, when it is turning and crossing uneven surfaces or obstacles.Articulated steering vehicles have two separate bodies connected by a steering hinge, which reduces the turning radius and enhances the load-bearing capacity. This mechanism also gives the vehicle enough room to install tires with large diameter and wide tread to save its cost and improve its trafficability. With these features, the articulated steering vehicles (e.g., articulated wheel loaders, scrapers and articulated trucks) are widely used in agriculture and construction industries. However, they are prone to rollover accidents due to their articulated steering mechanism and high center of gravity [1].To prevent rollover accidents, it is necessary to design a successful active safety warning system to notice any potential rollover danger. In other words, the meaningful and accurate rollover risk assessment is an important step in developing rollover avoidance schemes. Some existing studies addressed the risk assessment issue of front wheel steering vehicles [2][3][4], in which a dynamics model of front wheel steering vehicles was first built and the lateral stability of the vehicles was then quantitatively described for the rollover risk assessment. The same method has been adopted in the articulated steering vehicle.Due to the articulated mechanism, an articulated steering vehicle has higher degrees-offreedom (DOF) and requires higher dimensional modelling to cover the dynamic characteristics. Li et al.[5] built a 7 degrees-of-freedom (DOF) nonlinear dynamics model and Chi et al. [6] even built a 12-DOF model. These models describe the dynamic characteristics of articulated steering vehicles well and are only suitable for theoretical research, as their calculation process is relatively complicated and is not suitable for real-time calculations.On the other hand, some simplified models were built to explore the horizontal movement [7], steering instability [8] and snaking behavior [9] of articulated steering vehicles. These simplified models do not require complicated calculations. However, they can only characte...

An unscented Kalman filter based velocity estimation method for articulated steering vehicle using a novel dynamic model

“…The concrete details and parameters of the ADAMS model are shown in Appendix 2 and Table 1, which come from a home-built 1:4 ASV prototype. 29 The proposed dynamic model is compared with the ADAMS model, while the estimator is verified under noise input. Figure 4 shows the flow diagram of the simulation process.…”

“…Some high-dimensional ASV dynamic models without steering systems 27,28 only serve the purpose of theoretically analysing the dynamics, with no extension to the design of controlling and estimating methods for ASV. 29 To simplify the calculation, some research for lateral dynamic performance equivalents the articulated steering to a linear torsion spring and damper. 1,30 Other research further simplifies the steering system as an actor, 15,31 which assumes that the steering torque of the articulated steering system is known or can be directly measured.…”

Analytic Solutions for Wheeled Mobile Manipulator Supporting Forces

“…A recent study assessing the rollover risk in a 4-wheeled articulated-steering vehicle recognized the case in which one wheel off the ground is the dangerous state [30]. The authors used neural networks to estimate the rollover risk probability without detailed dynamics modelling, exploiting experimental data.…”

Driving Intention Inference Based on a Deep Neural Network with Dropout Regularization from Adhesion Coefficients in Active Collision Avoidance Control Systems

“…A multi-stream Convolutional Neural Network (CNN) was employed to extract multi-scale features by filtering images with receptive fields of different kernel sizes and the final decision is generated with different fusion strategies for driving behavior recognition [21]. A novel data-driven modelling methodology was proposed for the lateral stability description of articulated steering vehicles and a Recurrent Neural Network (RNN) model was built to accurately quantify vehicle lateral stability [22]. A long short-term memory (LSTM) network, which is a time cycle neural network, was used to evaluate real-time bus riding comfort and provide driving suggestions [23].…”