3URFHHGLQJVRIWKHWK0HGLWHUUDQHDQ&RQIHUHQFHRQ 7 &RQWURO $XWRPDWLRQ-XO\$WKHQV*UHHFH Embedded Electronic Control System for Continuous Self-Tuning of Motorcycle Suspension Preload F. Baronti*, F. Lenzi*, R. Roncella*, R.Saletti* and O. DiTanna** * University of Pisa, Dipartimento di Ingegneria dell’Informazione, Via Caruso 16, 56126 Pisa, (Italy) ** Piaggio S.p.A., R&D Department, Via Rinaldo Piaggio, Pontedera, (Italy) Abstract—Comfort and safety of vehicles significantly electronically-controlled actuator added to the spring and depends on the behavior of the suspension system. This is the damper. The actuator is computer-controlled in order particularly true in two-wheel vehicles where the to vary the adjustable suspension parameters and to equilibrium is fundamental. Variations of the static load on stabilize the vehicle, according to its state, as monitored the vehicle determine a compression of the suspension by sensors. spring that modifies the static equilibrium point with Such solutions are very efficient as far as performance respect to the optimal value. The system we describe is is concerned, but suffer the drawback of complexity and capable of continuously correcting the suspension preload of cost. These drawbacks are more stringent if we refer to a motorcycle without user intervention, in order to two-wheel vehicle applications. In this case, the compensate the load variations. The electronic system is requirements in terms of cost and robustness are very based on a microcontroller and a linear position sensor that tight. Moreover, the dynamics and the weight of the measures the suspension stroke. It executes a closed-loop involved vehicles are by far different with respect to four- control algorithm that adjusts the preload and maintains wheels, so that solutions feasible and affordable on a car the average value of the suspension stroke constant. The may not even be conceived on a motorcycle. Suspensions experimental results coming from road tests performed on a scooter are reported and discussed. equipped with electronic control are today available only on a few top-class motorcycles. The control is usually semi-automatic and the users are asked to select the I. INTRODUCTION preferred setting. Lower-class vehicles use suspensions equipped with a mechanical regulation of the preload, Suspensions are crucial subsystems of road vehicles obtained with a rotating gear. The regulation procedure is from both safety and performance points of view. They in principle simple, but rather tricky if manually applied absorb the shocks coming from the road asperities by any time the load on the vehicle changes. guaranteeing the road holding of the vehicle as well as the comfort of the occupants. Fundamental components Indeed, load variations are very important in of passive suspensions are the spring (the elastic element) motorcycles because the presence of a passenger and/or and the shock absorber (the damping element) [1]. baggage may increase the sprung masses above 100% compared to a driver-only load condition. It is obvious The suspension has to operate in many different road that suspension control becomes crucial also in medium and load conditions. Several mechanical settings are thus and low-class motorcycles, when the demand of available to improve its behavior, such as compression increased safety and comfort is raised from the users. In and rebound damping regulation and preload adjustment. any case, the suspension control should be automated and Adjustable suspensions were introduced many years achieved with low-cost systems affordable for two-wheel ago in the automotive field, as hydro-pneumatic systems. vehicles [2]. The many improvements over the years have also led to The aim of this work is to provide a motorcycle with a the implementation of active suspensions. An active or low-cost embedded electronic control system capable of intelligent suspension, as it can be defined, consists of an achieving the automatic self-tuning of the suspension preload, in order to overcome the limit of manual and Load C Load semi-automatic regulations. FP C The electronic embedded system described here P realizes a closed-loop control. It uses a linear position C L sensor to measure the suspension stroke and drives the suspension actuator, an electric DC motor that pumps oil in the hydraulic preload circuit, so that the average suspension stroke is maintained constant. The system is based on a microcontroller that implements the control algorithm without any intervention from the user. Two are the main problems to face with when Fig. 1 Preload effect on the height of the vehicle. designing such a control system: the hysteresis in the suspension response and the dependence of the measured 3URFHHGLQJVRIWKHWK0HGLWHUUDQHDQ&RQIHUHQFHRQ 7 &RQWURO $XWRPDWLRQ-XO\$WKHQV*UHHFH ⋅ = ⋅ ()+ M g k CL P (2) CL (mm) Preload M ⋅ g C = − P (3) L k CL* where CL is the static compression due to a mass M that loads the spring, g is the gravitational acceleration and k is the constant of elasticity of the spring. Controlling P Load (kg) makes possible to compensate the compression due the W1 W2 W3 mass M, to set the vehicle height with respect to ground and restore it to the optimal value, as indicated by the Fig. 2 Compression as function of load at given preload values. suspension manufacturer. Thus, the vehicle maintains the correct trim that would otherwise be altered by the suspension stroke on the vehicle dynamics [1]. This additional mass. means that the control algorithm has to manage the Fig. 2 shows the relationship between compression and negative effects due to hysteresis [3] [4] [5] [6] and to load with 3 different values of the actual preload. The extract static values from the measured data, in order to diagram also shows the hysteretic behavior of the spring. change the preload setting only when a static load change In any case, the compression can be maintained at the occurs, being affected as less as possible by the constant value CL*, if the preload is increased when the movements of the vehicle [7]. static load increases from W1 to W2 or W3. A Piaggio Beverly 500 scooter has been equipped with the system prototype and some road tests have been III. HARDWARE carried out. They show that the suspension is controlled by the electronic system in such a way to automatically A. System Architecture adapt its preload to any variation of the load conditions. The architecture of the system is shown in Fig. 3. The suspension is provided with preload regulation, driven by II. SUSPENSION PRELOAD a DC motor. The on-board Electronic Control Unit (ECU) Suspensions are designed to work when compressed. continuously monitors the stroke of the suspension (CL) The spring is thus pre-compressed when assembled. This by means of a linear position potentiometric sensor. The compression is called preload (Fig. 1). Let us express the actuator automatically adjusts the preload CP when CL preload value as the difference between the actual stroke differs from the optimal value, in order to compensate the of the spring and its value when no load is applied. In static mass variations of the vehicle due to a change of preload-adjustable suspensions, the preload value consists configuration. Preload may be adjusted in motion, but the of a fixed portion CFP, and a portion CP that is variable, vehicle dynamics affects the suspension compression in so that the actual preload P is: this case. Therefore the knowledge of the vehicle velocity is fundamental to take into account this effect. Finally, P = C + C (1) the lateral stand open (i.e. parked or unloaded vehicle) FP P should prevent the control algorithm to operate. Preload adjustment is very important in a motorcycle, The ECU is provided with a CAN bus transceiver and because it allows the compensation of static load a wireless link for communications. The CAN protocol variations of the vehicle, as suggested by the equilibrium compatibility has been provided to communicate with the equation applied to the suspension. In fact, internal vehicle network, if any, and to access to the messages available on the CAN network [8]. The wireless Linear Position Sensor Position Linear Fig. 3 System’s Architecture. 3URFHHGLQJVRIWKHWK0HGLWHUUDQHDQ&RQIHUHQFHRQ 7 &RQWURO $XWRPDWLRQ-XO\$WKHQV*UHHFH Analog C. Sensors and Actuator CAN Front-end The sensors used in the system provide the information Motor needed to execute the control algorithm. The vehicle Driver speed is measured by means of a Hall sensor, the output RS 232 µController of which is a square waveform with frequency Bluetooth proportional to the speed of the vehicle. The suspension static compression CL is measured by Power Supply means of a linear position sensor Penny and Giles MLS 130 [12] which is mounted in parallel to the suspension Fig. 4 ECU Architecture. spring. It is a potentiometric sensor for automotive applications, the output of which is the partition of the voltage reference VREF. that is applied to the link (Bluetooth protocol) is very useful during debug and microcontroller ADC input. This sensor is robust enough test, since it allows easy data logging on a remote for the application and represents a good trade-off notebook PC or PDA, between cost and performance. The measurement of the preload value CP is carried out B. ECU Architecture by means of a revolution counter integrated in the DC Fig. 4 shows the block scheme of the electronic board. motor of the suspension actuator, realized with a Hall The core is an 8-bit microcontroller from the Atmel sensor. No information about the direction of rotation is 80C51 family [9], with 8 ADC channels, 32 kB flash available. memory and CAN controller. We have used a Finally a digital signal provides the information on the microcontroller with redundant resources for this status of the lateral stand (open-closed).