Keywords

1 Differentiation to Current Force Feedback Actuator Designs

Current force feedback actuators which will be released in series production vehicles soon are using the same basic technology as column assisted electric steering systems. A small e-motor generates feedback torque with a worm gear towards the driver. A torque sensor is still required for a consistent steering feel to compensate friction tolerances from the manufacturing or wear [1]. A worm gear is a cost effective, simple and well-known component, the risk of a failing steer by wire system motivates the suppliers to use proven technology. Not all suppliers are focusing on this approach, e.g. Thyssenkrupp introduces a direct drive force feedback technology without a torque sensor [2]. Direct drive e-motors promise a rigid steering feel with the possibility to generate frequencies up to 50 Hz. The drawback which prevents a more common use of direct drive e-motors is the requirement for the end stop torque which is typically higher than 20 Nm. The exact number is still to be discussed among the OEMs. This would result in heavy and large e-motors which are cost intensive and have major drawbacks in terms of packaging. Figure 1 shows close to series production SbW columns from Nexteer Automotive and Thyssenkrupp.

Fig. 1.
figure 1

SbW Force Feedback Actuators Nexteer Automotive [3] and Thyssenkrupp [2]

Full size image

There are several possible solutions to using a small direct drive e-motor for functions related to steering feel but still providing high torque output:

  • Use the maxed out peak torque of the e-motor

  • Mechanical end stops in series to an e-motor

  • Mechanical brake in series to an e-motor

The different approaches shows there is no simple solution for the direct drive force feedback actuator in existence. A combination of an e-motor with an MR-brake described in the following could solve the tradeoff between good steering feel and reduced costs or packaging.

2 Design and Theoretical Benefits of the Actuator Combination

A magnetorheological brake is a device that uses the properties of a special fluid or powder to control the braking force. An electric coil is attached to either the rotor or the stator, and can generate a magnetic field when a current is applied to it. The magnetic field closes from the rotor to the stator through the gap, and it causes the MRP particles to align and form magnetic bridges between the two surfaces. Figure 2 shows a simulation without and with a magnetic field.

Fig. 2.
figure 2

Fraunhofer IWM)

Simulation of particle chains in an magnetic field [4] (©

Full size image

The magnetorheological brake is a simple, reliable, and efficient way of braking. To reach a counter torque of 25 Nm the MR-brake prototype consumes 60 W electrical power at 12 V [5]. An electric motor wheelbase such the Simucube 2 Pro generates a maximum torque of 25 Nm at 450 Watts peak [6].

Figure 3 shows on the left side the INVENTUS SbW actuator concept dimensions are significantly smaller than a compared gaming wheelbase. This hybrid force feedback actuator includes a 5 Nm e-motor and a 25 Nm MR-brake.

Fig. 3.
figure 3

INVENTUS SbW Force Feedback Actuator Concept Dimensions [5] vs. Simucube 2 Pro [6] (to scale)

Full size image

The major question is wether a hybrid force feedback actuator can generate the same quality of steering feel as a pure e-motor. In order to answer that, the hybrid actuator designed for this study (Fig. 4) did not focus on size or packaging. The e-motor itself can generate up to 15.7 Nm of continous torque and 30.0 Nm of peak torque. This enables a direct comparison on one system between the hybrid design and the pure e-motor with a disabled MR-brake. If both actuators are enabled, the e-motor torque can be limited to any low peak torque from a close to series production concept.

The actuator further has a MR-brake designed by INVENTUS Development GmbH with a peak torque of 8 Nm. Figure 4 shows the hybrid actuator cross-section and the real actuator including a high precision angle sensor.

Fig. 4.
figure 4

SbW Force Feedback Actuator with Direct Drive E-Motor and MR-Brake

Full size image

3 Controller and Torque Splitter

An MR-brake can only be controlled semi-actively and can therefore only take over two of four quadrants in the steering wheel torque- steering wheel angle velocity diagram during steering. An electric motor is required for e.g. active return. To design a control system for distributing the torque with good steering feel, the control must be advanced. This includes blending of actuator torques in the steering off-center and more e-motor proportion in the on-center region, thereby moving away from the ideal areas. Sudden changes in e.g. SWV may lead to unwanted behaviors due to settling time and minor differences in actual torque. To realize blending, factors between 0 and 1 were introduced and multiplied with the respectable torque command value of MR-brake and e-motor. The factors are realized via a characteristic field as shown in Fig. 5. At lower SWVs, the MR-brake factor is 0 and reaches a factor of 1 at a given SWV, called transition speed. A full takeover of torque from the brake only happens in areas where the MR-brake factor is 1 (resulting in an e-motor factor of 0). In areas where the MR-brake is active, but the factor is smaller than 1, the e-motor is partially taking over the torque in low SWV section, achieving torque blending of each actuator torques.

Fig. 5.
figure 5

Torque factor map of MR-brake for torque splitter

Full size image

For end stop behavior, a similar logic is used with two main differences. Firstly, since the achieved SWVs are low compared to driving even when pushing over the end stop torque limits, the transition speeds of the characteristic field are kept low, resembling more the ideal torque splitter. Secondly, the e-motor is not disabled in areas where the MR-brake is active to avoid fast on and off switching of e-motor torque due to the steeper map. Steering into the end stops results in an added torque of e-motor and MR-brake, steering away from the end stops is only controlled by the e-motor.

4 Simulation and Driver in the Loop Results

For the execution of Driver-in-the-Loop (DiL) tests, the hybrid actuator was implemented within a static driving simulator stationed in the laboratory of the Munich University of Applied Sciences. The control unit for the MR-brake and the inverter for the e-motor are connected to an embedded computer. This embedded computer handles the torque splitter and simulates a Pfeffer steering model [8].

To evaluate the functionality of the torque splitter, a Weave Test, 100 km/h, 30 °, 0.2 Hz, 4 m/s2 and End Stop Tests are performed. For both scenarios the actual torque values were computed utilizing the currents and their respective torque constants. It’s important to note that the setup did not incorporate a torque sensor. The graph in Fig. 6 shows the torques aginst the SWA for the Weave Test. The MR-brake becomes active during turning the steering wheel into the sine wave according to the torque-velocity diagram. Arriving at the peak of SWA sine wave, the MR-brake torque ramps down due to decreasing SWV. The e-motor takes over completely to generate the counter torque and a smooth transition can be seen. By superimposing the torques of each actuator, a well-defined hysteresis can be generated.

Fig. 6.
figure 6

E-motor and MR-brake torque splitting for Weave Test

Full size image

The graph in Fig. 7 shows the torques of the actuators over the measurement time, for the End Stop Test. The MR-brake becomes already active during slowly turning the steering wheel towards the end stop and the e-motor is inactive. At 4.5 s, when the end stop limit is reached the MR-brake and the e-motor torques are rising to the parametrized maximum. At 11 s during steering back to the center the e-motor torque is reduced slowly as well. The end stop transitions can be generated as a sum of the actuators or independently as shown in the measurement.

An expert evaluation of both scenarios in a DiL environment shows good subjective results, the transitions between the actuators where hardly noticeable, only the end stops where described as “little sticky”. The MR-brake is not released early enough, because the driver’s intention is sensed after oversteering the MR-brake torque.

Fig. 7.
figure 7

E-motor and MR-brake torque splitting for End Stop Test

Full size image

5 Summary and Outlook

In this study the potential advantages for a force feedback actuator with direct drive e-motor and MR-brake are presented. The design is cost effective, delivers a high maximum torque at low energy consumption and ensures flexible end stops. The steering feel of the design was analyzed objectively and evaluated subjectively by experts. The results show that the developed torque splitter feels smooth in the transitions, but the MR-brake tends to stick in the end stops. The reason for the behavior was identified in the angle sensor position at the end of the actuators. The sensor type will be changed towards a hollow shaft type and positioned between the steering wheel and the e-motor to improve the sticky feel. The torque input from the driver is measured early to release the MR-brake instead of first oversteer it. A second demonstrator with a series production EPS-motor including an updated MR-brake is currently being built up to further investigate the potential of the presented force feedback actuator with direct drive e-motor and MR-brake.