Makoto Yamakado

Improvement in Vehicle Agility and Stability by G-Vectoring Control

We extracted a trade-off strategy between longitudinal traction/braking force and cornering force by using jerk information through observing an expert drivers voluntary braking and turning action. Using the expert drivers strategy, we developed a new control concept, called ‘G-Vectoring control’, which is an automatic longitudinal acceleration control (No DYC) in accordance with the vehicles lateral jerk caused by the drivers steering manoeuvres. With the control, the direction of synthetic acceleration (G) changes seamlessly (i.e. vectoring). The improvements in vehicle agility and stability were evaluated by theoretical analysis and through computer simulation. We then introduced a ‘G-Vectoring’ equipped test vehicle realised by brake-by-wire technology and executed a detailed examination on a test track. We have confirmed that the vehicle motion in view of both handling and ride quality has improved dramatically.

Abnormal combustion detection and vibration reduction system

An apparatus and a method for detecting abnormal combustion in an internal combustion engine and an apparatus and a method for controlling vibration caused by abnormal combustion. The apparatus for detecting abnormal combustion includes a unit for detecting the rotation of an internal combustion engine, a unit for calculating rotation angular acceleration on the basis of the detected rotation, a unit for comparing the rotation angular acceleration at the position of a predetermined crank angle in one cycle of the operation of the internal combustion engine with a predetermined reference value, and a unit for judging from the comparison result whether abnormal combustion occurs in the internal combustion engine. The apparatus for reducing vibration caused by abnormal combustion offsets vibration caused by abnormal combustion and vibration of an electric generator to each other by passing a pulse-like field current in the electric generator in response to detection of abnormal combustion to generate vibration torque in the electric generator.

An experimentally confirmed driver longitudinal acceleration control model combined with vehicle lateral motion

Some approaches to emulating an expert drivers longitudinal and lateral integrated control using jerk information are studied in this paper. A trade-off strategy between longitudinal traction versus cornering traction was extracted experimentally by using jerk information to observe voluntary braking and turning during driving. The strategy is realised by determining the longitudinal acceleration command during combination motion by vehicle lateral jerk. The vehicle dynamical rationality of the method is also evaluated by computer simulation in view of instantaneous handling characteristics. Applying the strategy, we made a longitudinal driver model that works in coordination with any steering input. Adopting a programmed handling model and a preview-follower model for the steering input, we carried out computer simulation. We also carried out cornering and braking lane-change experiments with an expert driver. The simulation results and experimental results agree well and we confirm that the model can emulate a certain part of the expert drivers control strategy for trading off longitudinal and lateral accelerations.

A hybrid stability-control system: combining direct-yaw-moment control and G-Vectoring Control

In this study, a ‘hybrid stability-control’ system based on two concepts – G-Vectoring Control (GVC) and direct-yaw-moment control (DYC) – was developed. This system controls deceleration according to the information on vehicle lateral jerk and yaw moment according to the information on vehicle sideslip. It reduces the tendency to understeer by applying deceleration via GVC and reduces the tendency to oversteer by adding yaw moment via DYC. The tests with a vehicle fitted with this new GVC/DYC hybrid control confirmed that understeer can be reduced significantly more than that possible with conventional DYC only. It is concluded that this greater understeer reduction is a result of GVC preventing understeer prior to the skidding of the vehicle.

A yaw-moment control method based on a vehicle's lateral jerk information

Previously, a new control concept called ‘G-vectoring control (GVC)’ to improve vehicle agility and stability was developed. GVC is an automatic longitudinal acceleration control method that responds to vehicle lateral jerk caused by the drivers steering manoeuvres. In this paper, a new yaw-moment control method, which generates a stabilising moment during the GVC command and has positive acceleration value and the drivers accelerator pedal input is zero, was proposed. A new hybrid control, which comprises GVC, electric stability control and this new control, was constructed, and it was installed in a test vehicle and tested on a snowy surface. The very high potential for improvement in both agility and stability was confirmed.

Comparison and combination of direct yaw-moment control and G-Vectoring control

Previously, we developed a new control concept called ‘G-Vectoring control (GVC)’ to improve vehicle agility and stability. GVC is an automatic longitudinal acceleration control method that responds to vehicle lateral jerk caused by a drivers steering manoeuvres. In this paper, we compare GVC with the well-known direct yaw-moment control (DYC) method for daily driving ranges in particular. GVC shows good manoeuvrability (i.e. enhances both yawing and lateral acceleration) performance while maintaining a natural yaw and roll feeling in the early stage of cornering, while DYC is effective in improving vehicle stability during large lateral motions. We also re-review our previously proposed ‘hybrid control’ method that combines the strengths of GVC and DYC. A side-by-side test carried out for the hybrid control method and DYC only confirms the effectiveness of the hybrid control method.

Proposal of the longitudinal driver model in coordination with vehicle lateral motion based upon jerk information

Some approaches to emulating expert driverpsilas longitudinal and lateral integrated control using jerk information are studied in this paper. We proposed a longitudinal driver model which works in coordination with any steering input. The longitudinal acceleration command is calculated by the vehicle lateral jerk information. Adopting programmed and preview-follower model for the steering input, we compared with the expert drivers cornering and braking lane change test results both on dry asphalt and slippery road. The both results agree well, and we confirmed the model can emulate a certain respect of expert driverpsilas control strategy for trade-off between longitudinal and lateral acceleration.