Flight force measurements for a micromechanical flying insect

Key to the success of the micromechanical flying insect project is developing sensors for flight force measurement. At the lowest level of MFI control is the wing control system which will rely on wing and thorax mounted force sensors. These sensors will have the dual function of stroke by stroke force characterization and system identification as well as use in feedback control for all levels. There are two methods for force sensing on a flying robotic insect, measurements directly on the thorax, and body force measurement with a tradeoff in design between sensor bandwidth and sensitivity. 1.0 Introduction The Micormechanical Flying Insect (MFI) project [6], [13] aims to create a robotic insect, with a wingspan of 25mm, capable of sustained autonomous flight. Yan, et al suggested in [13] that force feedback would be both feasible and necessary for control in a micromechanical flying insect. It has been shown by Nalbach in [9] that insects such as Calliphora use sophisticated means to measure forces and torques for complicated guidance systems. The MFI will use a hierarchical control/sensor system, broken into the higher level of mission control down to individual wing control. The MFI is very biomimetic in nature, and the methods of force/torque sensing which are used are no exception. For autonomous strain sensing, the MFI will use a system very similar to insect campaniform sensilla (see [3]). Traditionally, measuring forces on a flying insect is performed by fixing the insect to a cantilever and optically measuring its position, or by measuring air flow vectors using particle image velocimetry (PIV) [7]. These methods will work with the MFI, however there are two reasons for avoiding it. First, it is desired to measure the forces autonomously during flight, where optical force reading would not be possible, and PIV would be difficult. Second, the nature of the forces generated by the MFI make it only necessary to measure certain torques, as will be discussed in section 2. Thus the MFI is best suited for either direct wing force measurement or body force measurement by way of cantilever based strain sensors. Each method of force sensing described in this paper has associated benefits and difficulties. Measuring the forces on generated by the MFI using off boar, or body force sensors give the benefits high sensitivity, and easy of use. Measuring the forces directly on the wing spar gives the benefit of being useable at the final scale, while providing difficulty in construction. Finally, placing strain sensors directly on the actuator makes for easier installation, allows for system identification to be performed, gives the ability to sense position, and can be used to protect the actuator from high strains. The main concern in the design of force sensors for microrobotics is sensitivity and bandwidth along with small size and low mass. What is unique about the MFI is the need for both high sensitivity due to small forces, and high bandwidth due to high wing beat frequency. The MFI currently uses folding processes described by Shimada in [12]. The goal of obtaining flight force measurements was not only to design high sensitivity, high bandwidth sensors for the MFI, but to integrate them into the current process. This paper describes three methods of measuring the forces generated by autonomous flying robotic insects, and gives preliminary results from tests done on the MFI thorax at the University of California, Berkeley. 2.0 Forces Acting on the MFI Most flying insects, such as the blowfly Calliphora, have control over three degrees of freedom (DOF) in their wing motion. These three degrees of freedom consist of the wing angle, angle of attack, and out of stroke plane deviation. Of these three, the MFI will use only the two most vital; wing position and angle of attack. These motions are achieved by using piezo-electric actuators, driving a mechanical amplifying thorax structure [6]. The free ends of the actuators are connected to slider crank mechanisms which convert the linear motion of the actuators into a rotation at the base of the fourbar. The thorax output drives a differential mechanism attached to a rigid wing. From the two one degree of freedom four-bars, the wing differential can produce wing motions consisting of the desired flapping and rotations. Flapping occurs when the actuators are in phase, and rotation occurs when the two are out of phase. Figure 1 shows the wing transmission system consisting of two actuators, two fourbars, the wing differential, and the wing. Figure 1: 2 DOF thorax and wing structure The MFI in free flight has six degrees of freedom, and has the use of two – two DOF wing structures to produce force/torque vectors in each of these six DOF’s. Thus measuring the generated wing forces is a crucial step towards control of the MFI as a whole. Figure 2: Insect coordinate system Section 3 describes a method of measuring the forces on the wing by placing sensors on the wing spars, thus the force is measured directly. Section 4 describes a method of measuring the total lift and drag by placing the MFI on a platform which measures the forces generated by the body. Finally, section 5 discusses placing sensors on the actuators as an alternative to placing gages on the wing. 3.0 Direct Force Measurement Directly measuring the flight forces involves measuring the moments on the wing using strain gages mounted directly in the wing spars. This has been achieved by using semiconductor strain gages bonded directly to each wing spar. This technique has the advantage of being scalable and practical for the final size MFI, and has the downside of difficulty in construction as well as possible interference with overall performance due to its wires. The direct force measurement was achieved in two stages, first at the 5X scale, then at 1.3X scale using similar techniques. 3.1 Wing Spar Sensing Strain gages are passive structures which change resistance when in compression or tension, thus by placing the gages across an excitation voltage, it is a simple matter to measure strain. The strain on a material is a unit-less quantity defined by ε = ∆l/l where l is the length of the material and ∆l is the change in length of the material caused by compression or tension. In the case of a rigid body, such as a wing spar, measuring the strain gives the moment directly.