Airborne pod macro-energy laser measuring and illuminating system with energy self-adaptive adjustment

Abstract

According to the energy self-adaptive adjustment airborne pod large-energy laser measuring and illuminating system provided by the invention, a transmitting party receives a reflected light spot of a ground target and evaluates whether the signal-to-noise ratio of the light spot in the current environment meets the requirements of a receiving party on target indication and / or target distance measurement; the laser emission energy is automatically and dynamically adjusted according to the evaluation of the environment under the condition that the communication with the receiver is not needed, so that the working reliability, the indication accuracy and the working endurance in various environments are enhanced.

Description

Technical Field

[0001] The invention relates to the field of laser ranging and indication of an airborne pod of an unmanned aerial vehicle, and in particular to an airborne pod high-energy laser ranging and indication system with adaptive adjustment. Background Art

[0002] Laser guidance system has the advantages of high guidance accuracy and strong anti-interference ability, which makes it more and more widely valued. When laser guidance is carried out, laser is often needed to guide the target, so the laser measurement system is indispensable.

[0003] Laser photogrammetry includes laser ranging and laser indication functions: the drone's onboard pod emits high-power invisible lasers that contain coded information. After the laser is irradiated on an object on the ground, the reflected light is received by the onboard pod for ranging. At the same time, the reflected light is received as an indication signal by other drones or seekers as a means of target indication.

[0004] Due to the presence of background stray light in the wild, there may also be backscattered lasers in the air, which are all interference light noise for laser spot detection. The energy of laser received at a long distance is weak, and the signal-to-noise ratio is low at this time. Moreover, the interference to lasers in different regions, different climatic conditions, and different environmental factors are different. Although blindly increasing the power can adapt to various environmental conditions, it consumes a lot of energy, which is not conducive to the endurance of the drone, and the entire system can work for a short time. However, if the laser power is not enough, under certain environmental conditions, the signal-to-noise ratio of the received signal may be low, resulting in target indication errors. How to adaptively adjust the laser transmission power under drone flight conditions is an urgent problem to be solved.

[0005] Some people have proposed using an additional laser test system to perform environmental testing. However, this is because the laser signals of different lasers are affected differently by the environment, and the laser signal of the light measurement system is encoded. It not only needs to receive light intensity information, but also needs to parse the encoding information. Therefore, the impact of the environment on it is different from that on general laser signals. For example, for general test laser signals, laser signals can be received in a certain environment, but in this environment, although the encoded laser signal of the same power can also receive the signal, the code cannot be accurately parsed, which will lead to errors in environmental assessment. This is one of the problems to be solved by the present invention.

[0006] In the existing technology, the UAV flight environment is usually estimated based on weather information, and the transmission power is adjusted according to the estimated parameters. However, it is obvious that weather and environmental information cannot be updated in real time, and there are many influencing factors, and it is difficult to accurately match the flight area. Therefore, this method is only feasible in theory, and the accuracy of practical application is low. Some people have also proposed to directly use laser signals to train neural networks to achieve energy adjustment, but this requires a more complex network structure and high computing power requirements, which is difficult to apply to UAV platforms. What's more difficult is that some adjustment systems require additional hardware structures, such as additional transmitting and receiving devices, which not only brings about a significant increase in energy consumption, but also affects the UAV payload, and requires a major transformation of existing UAVs to achieve, which is not suitable for rapid promotion.

[0007] Therefore, how to adjust the laser power in the existing UAV airborne pod high-energy laser imaging system in a low-cost, fast and accurate manner according to the actual flight environment to enhance the reliability and endurance of target ranging and indication is an urgent problem to be solved. Summary of the invention

[0008] In order to solve the problem, the present invention proposes a method and system for autonomously adjusting the laser coding transmission power in a laser measurement system. The transmitter receives the reflected light spot of the ground target, evaluates whether the signal-to-noise ratio of the light spot in the current environment meets the receiver's requirements for target indication, and autonomously and dynamically adjusts the laser transmission power without the need to communicate with the receiver, thereby enhancing the reliability of working in various environments.

[0009] An airborne pod high-energy laser imaging system with adaptive energy adjustment includes a laser transmitter and receiver, a processor, and a steering device.

[0010] The onboard pod laser transceiver emits laser signals to the preset target , and receive the reflected laser signal The processor estimates the ideal reflected signal based on the transmitted laser signal. , and run the following method at the same time:

[0011] Step 1: Build a hierarchical model:

[0012]

[0013] represents the linear kernel function, is the corresponding linear bias parameter, represents a nonlinear activation function, is the coordinate value. is the coordinate value; is the local eigenvector of the reflected signal;

[0014]

[0015] is a linear mapping, is the output linear vector, is the corresponding linear bias;

[0016] According to the output linear vector After the output layer operation, the illumination environment level is output ;

[0017] Step 2: According to the obtained lighting environment level, establish the regression model of signal quality and transmission power at each level:

[0018]

[0019] It is level The following regression model, is the ideal transmission power, that is, at level The minimum power at which the receiver can decode normally;

[0020] After the level model and regression model are trained, the optical transmitter receives the reflected signal, first estimates the level of the measured environment according to step 1, and then selects the appropriate regression model according to step 2 , calculate the current ideal transmission power, and adjust the transmission power of the laser transmitter and receiver according to the ideal transmission power.

[0021] Laser transceiver is used to transmit laser signals to the target and receive laser signals reflected by the target. The signal is used for target ranging, target indication, illumination level assessment, and laser energy regression prediction.

[0022] The laser signal is a coded signal.

[0023] The coding of the laser signal is time coding, space coding, or a mixed coding of time coding and space coding.

[0024] A processor, used to receive digital signals sent by a laser transceiver and to perform distance measurement according to an algorithm;

[0025] The processor is used to receive the user's instruction signal to drive the steering device, so that the steering device drives the laser transceiver to turn to the target and emit a laser signal to indicate the target.

[0026] The steering device is used to carry the laser transceiver and steer according to the processor driving signal to ensure that the laser signal emitted by the laser transceiver points to the target.

[0027] The steering gear is a mechanical steering gear or an optical steering gear.

[0028] A communication unit is also included for communicating with other drones or the ground.

[0029] An energy adaptive adjustment algorithm running on the airborne pod high-energy laser imaging system.

[0030] The invention and technical effects of the present invention are as follows:

[0031] 1. The present invention does not require additional hardware such as laser transmitters. The UAV airborne pod high-energy laser photometric system directly evaluates the laser emission and laser reception, and can directly evaluate whether the received laser meets the requirements. In particular, an environmental model is established, and on this basis, a neural network model dedicated to the UAV photometric system is designed to evaluate the environmental level. Without considering many detailed factors such as weather, light, and granularity one by one, the low computing power cost can achieve accurate measurement of the environmental level of the photometric system, and can accurately evaluate the impact of the current environment on the laser signal.

[0032] 2. Based on the accurate measurement of environmental levels, regression models are trained according to different levels to ensure the reliability and accuracy of the energy output of the lighting system in different scenarios, thereby improving the reliability of the entire system, effectively reducing energy consumption and extending battery life. DETAILED DESCRIPTION

[0033] The UAV flight environment will have a significant impact on laser photometry, such as particles in the atmosphere, ambient light, climate conditions, weather conditions, etc. There are many influencing factors. If we only start from these factors, we need to model each factor, which is a lot of work and difficult to be accurate. Moreover, the different degrees of influence of different factors in different flight environments will also lead to a decrease in the accuracy of the algorithm. To this end, the present invention innovatively proposes to directly use the high-energy laser photometry system of the UAV airborne pod to complete the photometry level assessment.

[0034] It should be pointed out here that the high-energy laser photometric system of the drone pod is a working system carried in the drone pod, which is mainly used for distance measurement and laser indication of targets. The present invention does not set up additional evaluation unit hardware and / or additional laser emission and laser receiving devices, because these additional devices will lead to inaccurate evaluation levels due to their inconsistency with the working parameters of the actual high-energy laser photometric system, affecting the laser power adjustment.

[0035] The high-energy laser imaging system of the drone's onboard pod includes a laser transmitter and receiver, a processor, and a steering device.

[0036] (1) Laser transceiver, used to transmit laser signals to the target and receive laser signals reflected by the target. The signals are used for target ranging, target indication, illumination level assessment, and laser energy regression prediction.

[0037] The laser signal is a coded signal, which makes it easy to distinguish signals from each other when multiple drones work together. The coding can be time coding, space coding, or a hybrid of time coding and space coding.

[0038] The laser transceiver converts the received laser signal into a digital signal and sends it to the processor.

[0039] (2) A processor, which is used to receive digital signals sent by a laser transceiver and can perform ranging according to an algorithm.

[0040] It can also receive user instruction signals to drive the steering device, so that the steering device drives the laser transceiver to turn to the target and emit laser signals for target indication.

[0041] It is also possible to evaluate the high-energy laser illumination level through an algorithm based on the received digital signal and to perform adaptive control of the laser transceiver power. The specific algorithm will be described in detail below.

[0042] (3) A steering device, which is used to carry the laser transceiver and steer according to the processor driving signal to ensure that the laser signal emitted by the laser transceiver points to the target. The steering device can be a mechanical steering device or an optical steering device.

[0043] In addition, the drone pod also includes a communication unit for communicating with other drones or the ground. The remaining drone components such as power module, flight control module, fuselage, remote control console, etc. are similar to existing drones and will not be described in detail.

[0044] The laser power adaptive adjustment algorithm of the UAV airborne pod high-energy laser imaging system runs in the processor, as follows:

[0045] Step 1: Assessment of the environmental level of the UAV airborne pod’s high-energy laser lighting.

[0046] The airborne pod laser transceiver emits laser to the preset target and receives the reflected light. Suppose the airborne pod’s transmission signal is , the received signal is .

[0047] Furthermore, both the transmitted signal and the received signal are two-dimensional digital signals, which can be expressed in matrices as follows:

[0048]

[0049] , Represents an element in the matrix, Indicates the position of an element in a matrix.

[0050] The environment model is established as follows:

[0051]

[0052] symbol represents the natural exponential function, symbol represents the natural logarithm function, symbol represents the convolution operation, is the convolution kernel function, Represents the position of the convolution kernel element.

[0053] Equation 1 is used to estimate the ideal reflected signal .

[0054] It is understood that in another embodiment, the ideal reflection signal can also be derived according to other existing theories. In another embodiment, the ideal reflection signal can also be obtained by measuring the optical transmitter and the optical receiver under preset conditions.

[0055] According to the transmission signal Calculate the ideal reflected signal , further, according to the ideal reflected signal , and receive signal , evaluate the impact parameters of the environment.

[0056] The evaluation model is established as follows.

[0057]

[0058] symbol Indicates taking the maximum value, symbol It means taking the sum of all values. represents the linear kernel function, is the corresponding linear bias parameter. Represents a non-linear activation function. is the coordinate value. is the coordinate value.

[0059]

[0060] Wherein arctan is the inverse tangent function, and min represents a small value. By using the above optimized excitation function, the environmental feature extraction can be made more accurate.

[0061] It is a description of the local characteristics of the two groups of reflection signals, thereby reflecting the environmental characteristics. By taking the maximum value to obtain the extreme value of the local characteristics of the reflection signal under ideal conditions, and then combining it with the mean value of the local characteristics of the real reflection signal, a feature quantity with high correlation with the environment can be obtained, thereby reflecting the environmental characteristics.

[0062] Further

[0063]

[0064] is a set of linear mappings that transform local features Mapping to linear vectors , thereby establishing the relationship characteristics between local features, which is another factor reflecting environmental characteristics. is the corresponding linear bias, is the activation function, is a vector Vector mark of .

[0065] By establishing the relationship between local features, the interference of environmental noise can be further reduced, and the resulting model recognition effect is better.

[0066] at last

[0067]

[0068] in

[0069]

[0070] is a set of linear mappings, is the corresponding linear bias, is the activation function, is a vector Vector mark of .

[0071] Output Indicates the measurement environment level assessment, for example, level 1-5, representing 5 different degrees of environmental impact.

[0072] This level is different from any existing environmental assessment model. It does not focus on specific environmental parameters, but directly measures the impact of the environment on the laser signal. It is also important to point out that this is specifically for the impact of the laser signal on the measurement system, not just any laser signal.

[0073] This is because if the laser signals of different lasers are affected differently by the environment, and the laser signal of the lighting measurement system is encoded, it is not only necessary to receive the light intensity information, but also to parse the encoded information. Therefore, the impact of the environment on it is different from that on general laser signals.

[0074] For example, for a general test laser signal, the laser signal can be received in a certain environment, but in this environment, although the coded laser signal with the same power can also be received, the code cannot be accurately parsed, which will lead to environmental assessment errors. This is one of the problems to be solved by the present invention.

[0075] It is understandable that multiple convolutional layers and / or pooling layers can be added to improve accuracy, but the above network layers are the most important. Under limited computing power, the above network layers of the present invention can already achieve good detection and evaluation results. If applied to a platform with higher computing power, one or more other network layers can be added as needed.

[0076] By establishing a training sample set and implementing the training process for the above models 2-6, the above output feature vectors can be established. The training can be implemented using the BP algorithm using the following cost function:

[0077]

[0078] is the label value of the training sample. is the cost calculated by cost function 7. By minimizing the cost function, the model output is approximated to the sample label, thereby completing the training of the recognition model.

[0079] After training, the model reflects the signal according to the ideal situation of the input , and receive signal , take the maximum value of the five output dimensions and determine that the current environment corresponds to an environment level.

[0080] Step 2: Regression method of laser energy under given illumination environment level.

[0081] According to the lighting environment level obtained in step 1, a regression model of signal quality and transmission power at each level is established.

[0082] First, according to the transmitted signal , and the received signal R, calculate the signal quality.

[0083]

[0084] The transmit power and signal quality follow the following regression model:

[0085]

[0086] It is level The regression model below, for example, uses a polynomial function, is the ideal transmission power, that is, at level The minimum power at which the receiver can decode normally.

[0087] Step 1 in the level The training data collected below is used to train the regression model .

[0088] In order to improve the training effect of the regression model, the cost function is adopted as follows:

[0089]

[0090] in is a training sample, is the ideal transmission power calculated according to the regression model, is its marked ideal transmission power. is an exponential function. is an empirical parameter, obtained based on experiments ~1.6, preferably 1.4. Compared with the classic model, the above cost function improves the training effect and is more robust to noise.

[0091] After the model training is completed, the transmitter receives the reflected signal, first estimates the level according to step 1, and then selects the appropriate regression model according to step 2 , calculate the current ideal transmit power, and adjust the transmit power according to the ideal transmit power.

[0092] The present invention proposes a method for autonomously adjusting the laser coding emission energy. In order to verify the algorithm effect, 15 common scenes of 5 categories (city, forest, wasteland, farmland, water area) were selected, and 6 measurements were performed on each scene. The 6 measurement times of each scene were randomly selected (the first day of each week), and the measurement results were averaged as shown in the following table.

[0093] Experimental results show that the proposed method is accurate in evaluation, with a low error between the predicted power and the actual ideal power, meeting the actual UAV application requirements (<10%).

[0094] Different scenarios The prediction error of the present invention Methods for adjusting the existing technology according to environmental parameters Conventional CNN neural network City 6.4% 23.5% 12.8% forest 3.7% 19.7% 10.5% wasteland 2.8% 22.3% 9.7% farmland 4.5% 30.6% 13.7% Waters 5.1% 28.8% 11.4%

[0095] It can be understood that all the above descriptions are only for those skilled in the art to better understand the technical solutions and technical effects of the present invention, but cannot be used as a limitation on the protection scope of the present invention.

Claims

1. An airborne pod high-energy laser imaging system with adaptive energy adjustment, characterized in that: Including laser transmitter and receiver, processor, and steering device, The onboard pod laser transceiver emits laser signals to the preset target , and receive the reflected laser signal The processor estimates the ideal reflected signal based on the transmitted laser signal. , and run the following method at the same time: Step 1: Build a hierarchical model: ， represents the linear kernel function, is the corresponding linear bias parameter, represents a nonlinear activation function, is the coordinate value. is the coordinate value; is the local eigenvector of the reflected signal; ， is a linear mapping, is the output linear vector, is the corresponding linear bias; According to the output linear vector After the output layer operation, the illumination environment level is output ; Step 2: According to the obtained lighting environment level, establish the regression model of signal quality and transmission power at each level: ， It is level The following regression model, is the ideal transmission power, that is, at level The minimum power at which the receiver can decode normally; After the level model and regression model are trained, the optical transmitter receives the reflected signal, first estimates the level of the measured environment according to step 1, and then selects the appropriate regression model according to step 2 , calculate the current ideal transmission power, and adjust the transmission power of the laser transmitter and receiver according to the ideal transmission power.

2. The system according to claim 1, characterized in that: Laser transceiver is used to transmit laser signals to the target and receive laser signals reflected by the target. The signal is used for target ranging, target indication, illumination level assessment, and laser energy regression prediction.

3. The system according to claim 2, characterized in that: The laser signal is a coded signal.

4. The system according to claim 3, characterized in that: The coding of the laser signal is time coding, space coding, or a mixed coding of time coding and space coding.

5. The system according to claim 1, characterized in that: The processor is used to receive the digital signal sent by the laser transceiver and can perform ranging according to the algorithm.

6. The system according to claim 1, characterized in that: The processor is used to receive the user's instruction signal to drive the steering device, so that the steering device drives the laser transceiver to turn to the target and emit a laser signal to indicate the target.

7. The system according to claim 1, characterized in that: The steering device is used to carry the laser transceiver and steer according to the processor driving signal to ensure that the laser signal emitted by the laser transceiver points to the target.

8. The system according to claim 1, characterized in that: The steering gear is a mechanical steering gear or an optical steering gear.

9. The system according to claim 1, characterized in that: A communication unit is also included for communicating with other drones or the ground.

10. An energy adaptive adjustment algorithm running on the airborne pod high-energy laser imaging system as described in claims 1-8.

Images