A self-sustaining wireless sensing and flight control device for beetles

Insects exhibit the highest level of taxonomic diversity within the animal kingdom, spanning sizes from less than 1 mm to more than 20 cm.¹ The remarkable variety of locomotor patterns displayed by insects, including flight, walking, swimming, jumping, and burrowing, reflects the enduring result of prolonged evolutionary processes.^2–4 These efficient and intricate functionalities endow insects with the ability to achieve widespread global dispersal.

In the past decade, researchers have pioneered the development of insect robots by integrating miniature computers with living insects, known as insect-electronics hybrid robots.^5–8 Desired behaviors were elicited by directly stimulating the insect brain, ganglia, or motor muscle tissues.⁹ Sato et al.¹⁰ demonstrated the controllability of beetle takeoff by applying alternating positive and negative pulses to the left and right optic lobes, with individual negative pulses applied to these lobes effectively terminating beetle flight. Application of positive pulses to the left and right basal muscles enabled the beetle to execute a lateral turn,¹¹ while stimulation of the third axillary (3Ax) muscles induced an ipsilateral turn.¹² Additionally, Li et al. designed a proportional-derivative feedback controller using a 3D motion capture system to apply appropriate frequency pulses to the basal and 3Ax muscles for maintaining the beetle’s flight altitude.¹³ 

Insect robots can be equipped with diverse sensor payloads tailored to specific requirements, enabling them to function as mobile sensing nodes for environmental information detection.¹⁴ For instance, Iyer et al.¹⁵ equipped a beetle with an image sensor and developed a microactuator to enable real-time image acquisition by controlling the camera direction. The system demonstrated an operational duration of up to 6 h. Iyer et al. equipped bees with miniature sensor payloads to record data on light intensity, humidity, and spatial position during their flights¹⁶ and attached an airdropping sensor payload to Manduca sexta moths for environmental monitoring purposes.¹⁷ Similarly, Nanyang Technological University developed a series of sensing devices specifically designed for cockroaches, enabling these insects to autonomously navigate complex terrains and accurately detect human locations.¹⁸ Xiong et al.¹⁹ designed a microphone array backpack that allowed the cockroach to effectively perceive sound sources. However, the sustainability of these systems is compromised in outdoor environments due to inefficient energy harvesting and management, as well as the limited capacity of miniature batteries.

Due to the miniaturization of wireless systems and the reduction in power consumption, utilizing environmental energy sources to directly power electronic devices or charge energy storage devices has become a prominent and recent research topic.^20–23 Iyer et al.²⁴ proposed a lightweight, seed-like structure that integrates temperature and humidity sensors, along with solar cell films, energy storage components, and energy harvesting circuits, enabling sustainable detection of temperature and humidity. Dias et al.²⁵ presented an autonomous multisensor system incorporating temperature and air humidity, powered by an energy harvester, thereby achieving a significantly extended operational duration from 136 h to over 266 h.

In this study, we propose a self-sustaining wireless sensing and flight control microelectronic device suitable for beetles. Section II provides a comprehensive overview of the microelectronic device design, including the energy management module and Bluetooth slave module. Section III presents an evaluation of the microelectronic device’s wireless flight control for beetles. Section IV is dedicated to testing the operational duration of the device.

The microelectronic device consists of a power management module, a Bluetooth slave module, and energy storage components. Additionally, it features wireless control capabilities for beetle flight, along with wireless vision and attitude angle sensing functionalities that enable real-time image and posture detection and remote wireless transmission, as shown in Fig. 1.

The microelectronic device comprises two modules: the power management module and the Bluetooth slave module, as depicted in Fig. 2. The power management module is responsible for energy harvesting and management, while the Bluetooth module enables wireless flight control and sensing capabilities.

The solar cells continuously harvested solar energy and converted it into electrical energy, which served as input to the power management module. A lithium battery was designated as the energy storage component of this module. The Bluetooth slave module obtained energy from the stored power in the power management module through a path established by an analog switch, while the power management chip controlled the on/off state of this path based on the lithium battery’s voltage. Once the lithium battery accumulated sufficient energy, the pathway was activated, allowing the Bluetooth slave to continue functioning. The Bluetooth slave received electrical stimulation commands sent by the Bluetooth host through the antenna to induce beetle flight and transmitted collected images and posture information back to the Bluetooth host. Subsequently, the Bluetooth host sent these images and pose information to LabVIEW on the personal computer (PC) via a universal serial bus (USB).

The microelectronic device has three operating states: shutdown, cold start, and normal operation. When illumination is insufficient and the solar cell cannot adequately charge the lithium battery, the device remains in shutdown mode, disabling wireless image sensing. As illumination increases, the device undergoes a cold start to allow the lithium battery to accumulate sufficient energy. During normal operation, visual sensing and wireless transmission are enabled. With ample sunlight, the solar cell can supply enough energy to the Bluetooth slave module, and the lithium battery can maintain a steady voltage for continuous sensing. Conversely, insufficient illumination requires the lithium battery to provide additional energy to the Bluetooth slave module, causing its voltage to drop until the device enters shutdown mode.

The main functions of the power management module are energy collection and energy management. We selected small solar thin-film cells (Powerfilm ONP1.2-12 × 24) as energy collection devices, which have dimensions of 12 × 24 mm² and a weight of 90 mg. These solar thin-film cells can provide 4 mW of power under sufficient sunlight, with an output voltage of 1.2 V and a current of 3.3 mA. To enhance energy collection, we connected three of these solar thin-film cells in series [Fig. 3(a)]. For energy management, we chose the BQ25570 energy management chip, which can collect small amounts of energy from the collectors and store it in lithium batteries [Fig. 3(b)]. The power management system integrated into the BQ25570 chip ensures efficient energy conversion and storage. To achieve automatic startup and shutdown without the need for microcontroller control, we selected an analog switch chip (TMUX1101) to establish an autonomous power supply circuit between the energy management module and the Bluetooth slave module [Fig. 3(c)]. The logic control input (SEL) of the analog switch is connected to the power indicator output pin (VBAT_OK) of the BQ25570, allowing the system to independently control the power supply path to the Bluetooth slave based on the energy storage status. The size of the designed energy management circuit board is 12 × 11 mm². We chose a 20 mAh small lithium battery (8 × 6.5 mm²) as the energy storage element [Fig. 3(d)], which provides energy to the device when the solar cell input is insufficient. The workflow of the power management module is illustrated in Fig. 3(e).

The existing wireless vision system required large batteries due to its high power consumption, which consequently increased the system’s size and weight. Therefore, a low-power wireless protocol was necessary to reduce battery capacity requirements. Considering both the bandwidth needed for image transmission (greater than 1 Mbps) and the demand for low power consumption, we selected the Nordic NRF52840 Bluetooth chip, which supports a bandwidth capacity of 2 Mbps. The NRF52840 chip wirelessly receives electrical stimulation signals, collects images, and records flight attitude data. To detect the flight posture of beetles, we integrated attitude sensors (MPU9250) into the Bluetooth slave module design. The MPU9250 combines accelerometers, gyroscopes, and magnetometers to provide precise attitude and motion tracking data. Additionally, it supports I2C and serial peripheral interface (SPI) communication interfaces for integration with the NRF52840.

For environmental image detection, we chose the Himax HM01B0 (1.1 mW, Quarter Quarter Video Graphics Array (QQVGA) resolution) low-power image sensor, which weighs only 180 mg. The Bluetooth slave communication design includes a spiral antenna made of copper wire, weighing 70 mg with a length of 9 mm, and a received signal power of 3 dB. Additionally, we designed an electric stimulator module on the Bluetooth slave. One end of the microelectrode connects to the electrical stimulation module, and the other end connects to the beetle’s flight muscle. The electrical stimulation module features two sets of channels connected to the left and right flight muscles. The design of the Bluetooth slave module is depicted in Fig. 4.

The effectiveness of beetle flight control is significantly influenced by the system’s weight as the maximum load capacity of the beetle is limited to 2 g. The weight analysis of the system components is presented in Table I, which reveals that the overall weight of the system amounts to 1.58 g. Therefore, beetles can carry this device for motion control.

The insect robot in this study was achieved through the surgical implantation of wire electrodes into the flight muscles of the Mecynorhina beetle. This species was selected based on its remarkable payload capacity, allowing it to bear loads ranging from 2 to 3 g.

1. Ethics approval: The Mecynorhina beetles do not belong to the protected insects prescribed by the Chinese government and can be traded freely. Southeast University only carries out the ethical review for experiments related to humans and animals, and insect-related experiments do not require ethical review.

2. Stimulation profile: Stimulation of the third axillary (3Ax) muscle generated lateral forces to induce flight maneuvers. The parameters for flight maneuver stimulation were set as follows: frequency was 50 Hz, voltage was 1.6 V, duty cycle was 20%, and duration was 200 ms.

3. Electrode insertion: To initiate the procedure, the beetle was anesthetized by placing it in a food-grade CO₂-filled plastic container for 90 s. Small incisions were then made in the exoskeleton using a medical blood collection needle with a diameter of 0.21 mm. For the electrodes, we used polyvinylidene fluoride (PVDF) coated thin silver wire with an 80 μm bare core. The electrodes were carefully inserted into the left and right 3Ax muscles to a depth of 3 mm. Finally, the device was securely attached to the posterior pronotum of the beetle to facilitate precise muscular stimulation and environmental sensing.

The flight posture of beetles can be represented by Euler angles: pitch angle, roll angle, and yaw angle, which describe rotation around the x, y, and z axes, respectively. The posture of beetles was detected using the MPU9250 inertial sensor. Additionally, the central processing unit generated pulse width modulation (PWM) square waves to stimulate the third axillary muscle of the beetles, inducing them to turn toward the stimulated side. By utilizing real-time yaw angle information, we were able to dynamically adjust their flight posture, resulting in directed control of beetle flight. Electrical stimulation of the 3Ax muscle causes beetles to turn toward the same side as the stimulation. The process of directional flight control in beetles is illustrated in Fig. 5.

The impact of turn stimulation on beetles in a state of free flight is depicted in Fig. 6. As shown in Figs. 6(a) and 6(b), following the stimulation of the right 3Ax muscle, the beetle consistently exhibited a turning motion toward the right side, with its yaw angle progressively increasing relative to the preset angle, indicating successful execution of the right-turn stimulation. The effect of beetle right-turn control can be seen in supplementary material, Movie S1.

The effect of mixed turn stimulation on the beetles is illustrated in Fig. 6(c). Sequential left-turn and right-turn stimulations were applied after the initiation of flight in the beetles. As shown in Fig. 6(d), the yaw angle of the beetles relative to the preset angle initially decreased and then increased, indicating successful steering through combined left-turn and right-turn stimulation. The control effect of beetle mixed turning stimulation can be seen in supplementary material, Movie S2.

After validating the efficacy of the turning stimuli, we conducted an experimental investigation to assess the success rates of left and right turning stimuli. The experiment involved two beetles executing ten left turn stimuli and ten right turn stimuli, respectively. A stimulus was considered successful if a significant turning response was observed following it. The experimental results showed that when the right turn stimulus command was given, the two beetles successfully turned right 15 times. When the left turn stimulus command was given, they successfully turned left 14 times. The two beetles conducted a total of 40 turning experiments, successfully turning 29 times, resulting in a turning success rate of 72.5%.

We also conducted experiments on regulating straight-line flight in beetles. When the directional flight control was activated, the real-time yaw angle measured by the MPU9250 was compared with the previously set yaw angle. If the difference exceeded 20°, an automated sequence of stimuli for left or right turns was initiated by the microelectronic device to adjust the beetle’s flight direction. The results of the directed flight control are illustrated in Figs. 7(a) and 7(b), depicting the effect diagram and the recorded yaw angle curve, respectively. It can be observed that the microelectronic device effectively rectified the beetle’s flight attitude. The control effect of the beetle’s straight-line flight can be seen in supplementary material, Movie S3.

Hovering flight control in beetles was achieved by continuously applying stimulating signals for either left or right turns. The stimulation interval was set to 50 ms, and the number of stimulation pulses was set to 10 via the computer interface, allowing us to repetitively execute this sequence for precise control over the beetle’s hovering flight. The resulting effect is illustrated in Fig. 7(c). The experiment on hovering flight also validated the capability of inducing sustained turning in beetles through continuous stimulation. The control effect of the beetle’s hovering flight can be seen in supplementary material, Movie S4.

The device design utilizes three solar cells connected in series as energy-harvesting devices, capable of producing a maximum power of 12 mW. Under sufficient lighting conditions, sustainable image sensing can be achieved if the device’s image sensing power consumption remains below 12 mW. First, we conducted power consumption tests on the device in three operational modes: capturing one image per second, performing only yaw angle acquisition, and capturing one image per second while performing yaw angle acquisition. The average current consumption of these three modes is illustrated in Fig. 8(a). Since the voltage of the input Bluetooth slave module remains stable at 3.3 V, the power consumption of the device in these three operating modes is 11.28, 15.15, and 23.63 mW, respectively. It is evident that, in the mode of capturing one image per second, the device consumes less power than the solar input power, enabling it to continue operating under sufficient illumination conditions. However, upon enabling the attitude sensor, the power consumed by the Bluetooth slave module exceeded the maximum power provided by the energy management module, resulting in the inability to sustain sensing operations.

We tested the device’s sustainable sensing performance in indoor environments using a lamp with a maximum illumination of 20 000 lux to simulate solar energy. The device was placed under the lamp, and its operating time was determined by monitoring the displayed images on the PC interface. As depicted in Figs. 8(b) and 8(c), under an indoor illuminance of 20 000 lux, the operational duration of the device increased from 4.83 to 7.08 h in the mode of capturing one image per second, representing a significant prolongation of 46.6%. Similarly, when yaw angle acquisition was incorporated along with capturing one image per second, the operational duration extended from 1.61 to 2.43 h, indicating a substantial increase of 50.9%. These results clearly demonstrate that the device’s sustainable working capability has significantly improved under indoor lighting levels of 20 000 lux.

We tested the performance of the micro wireless sensing device in outdoor environments under direct sunlight. To monitor the device’s operation, images were displayed on the PC interface. We used a digital ammeter and a digital voltmeter to measure the current and voltage of the solar thin-film cells and lithium batteries, respectively. Experiments were conducted in two outdoor operating modes: capturing one image per second and capturing one image per second while acquiring a yaw angle. The experiments commenced at 10:00 AM, and the maximum measured illuminance reached ∼150 000 lux. In the mode of capturing one image per second, under favorable illuminance conditions where the solar output current exceeded 3.3 mA, a gradual decrease in lithium battery voltage indicated balanced power consumption between solar output power and device consumption power. However, as illumination levels decreased, the voltage of the lithium battery experienced a rapid decline, leading to the device ceasing operation by 6:30 PM [Fig. 9(a)]. The device operated continuously in this mode for a duration of 8.6 h. When operating in the mode of capturing one image per second and acquiring a yaw angle, the power consumption of the device exceeded the output power of the solar cell. As a result, the operational time was reduced to 3.11 h. Nevertheless, this operational time was 92% longer than the discharge time observed without solar cells [1.61 h, Fig. 9(b)].

The ability of the device to restart and function is crucial for sustainable operation. Starting at 10:00 AM the next day, the device transitioned from shutdown (2.8 V) to restart (3.3 V) within 10 min, as observed from the image display on the PC interface [Fig. 9(c)]. The input power (13.2 mW) of the device exceeded the power consumption (7.956 mW) of the Bluetooth slave, resulting in a rapid increase in the voltage of the lithium battery. After ∼5 h (3:00 PM) of sufficient sunlight exposure, the voltage of the lithium battery reached 3.73 V. When the weather became cloudy and illumination decreased, although the output current of the solar cell reduced (3:50 PM), the voltage stored in the lithium battery was adequate to sustain system operation until nightfall (4.83 h). Therefore, it was demonstrated that sustainable operation could be achieved by capturing one image per second under favorable outdoor lighting conditions. Moreover, the system significantly extended its operational time by simultaneously capturing yaw angle and one image per second under optimal outdoor lighting conditions. The total weight of the wireless sensing and control device (1.58 g) is less than the maximum load capacity of the beetle (2 g), allowing the beetle to carry the device and move normally outdoors, as depicted in Fig. 9(d). The beetle was placed on the lawn to facilitate unrestricted crawling [Fig. 9(e)], and a real-time first-person perspective image of the beetle was wirelessly transmitted to the host, as shown in Fig. 9(f).

The NRF52840 chip we selected uses differential transmission to send RF input and output signals, which reduces RF signal scattering and minimizes interference from other signals. To achieve longer signal transmission distances, we can use high-gain Bluetooth host flat-panel antennas (12 dB) and place them as parallel as possible to the micro antenna on the same horizontal plane. This maximizes the sensing surface for wireless signal transmission and reception, thereby extending the wireless transmission distance. The wireless transmission distance of this device is ∼160 m.

In practical outdoor applications, the real beetle system offers advantages such as good mobility, the absence of complex mechanical structures, and strong anti-interference capabilities. Additionally, its insect-like appearance provides effective concealment in military detection. Real beetles carrying sensors can access dangerous or narrow areas that humans cannot reach and perform tasks such as environmental detection or disaster relief. We presented some monitoring images of the system during outdoor flight. Some photos of beetles carrying microelectronic devices for outdoor flight monitoring are shown in Fig. 9(g).

This paper proposes a self-sustaining wireless sensing and flight control microelectronic device weighing only 1.58 g. This device enables directional flight control by manipulating beetle turning behavior, achieving a success rate of 72.5% in turning control experiments. Additionally, the microelectronic device integrates image and attitude sensing functionalities. It utilizes solar cells for energy harvesting, enabling self-sustaining wireless sensing. Under indoor lighting conditions of 20 000 lux, the operational time of the device increased by 46.6% when capturing one image per second. Furthermore, incorporating yaw angle acquisition alongside image capture extended the operational duration by 50.9%. In favorable outdoor lighting, the device maintained continuous operation in the mode capturing one image per second, with an operational duration extended by 92% in the mode combining image and attitude sensing per second. Benefiting from insects’ excellent mobility and environmental adaptability, this microelectronic device can serve as a mobile sensing node by directly integrating energy harvesting devices and visual sensors onto insects. This approach not only prolongs the operational times of insect vision robots outdoors but also reduces reliance on batteries and the frequency of replacements, enhancing the feasibility of long-term outdoor deployments. Ultimately, our research provides insights into self-sustaining sensing microelectronic devices for long-term environmental monitoring and promotes their practical application in real-world scenarios.

In the future, we could explore the synergistic effects of multiple muscles and quantitatively analyze the behavioral responses of insects to electrical stimulation. To address the deficiencies in the electrical and mechanical properties of existing stimulation electrodes, we could also use highly biocompatible and electrically efficient flexible electrode materials to manufacture multi-point flexible stimulation microelectrodes. This would further improve the accuracy of electrical stimulation to achieve better performance.

The effect of beetle right-turn control can be seen in supplementary material, Movie S1. The control effect of beetle mixed turning stimulation can be seen in supplementary material, Movie S2. The control effect of the beetle’s straight-line flight can be seen in supplementary material, Movie S3. The control effect of the beetle’s hovering flight can be seen in supplementary material, Movie S4.

This work was supported by the Postgraduate Research and Practice Innovation Program of Jiangsu Province under Grant No. KYCX22_0290.

The authors have no conflicts to disclose.

Ethics approval for experiments reported in the submitted paper on animal or human subjects was granted. The Mecynorhina beetles do not belong to the protected insects prescribed by the Chinese government and can be traded freely. Southeast University only carries out the ethical review for experiments related to humans and animals, and insect-related experiments do not require ethical review.

Xin Huang: Conceptualization (equal); Formal analysis (equal); Writing – original draft (equal); Writing – review & editing (equal). Wenhao Zhao: Investigation (supporting); Writing – original draft (supporting). Meisong Yuan: Investigation (supporting). Kaixuan Sun: Investigation (equal). Bo Yang: Funding acquisition (equal); Resources (equal); Supervision (equal).

The data that support the findings of this study are available from the corresponding author upon reasonable request.