Environment information updating method and device and storage medium

Abstract

The invention discloses an environment information updating method and device and a storage medium. The environment information updating method comprises the steps that sensor data of operation equipment at the current moment and the global probability of each environment element in prior environment information at the current moment are acquired; based on the obtained sensor data of the operation equipment at the current moment, obtaining each target detected by each operation equipment, and calculating the local probability of each target; and calculating a transition probability between the target data of each target and each environmental element based on the calculated local probability of each target, and obtaining the target data of each target according to the obtained global probability of each environmental element in the prior environment information at the current moment and the calculated transition probability. And updating to obtain the global probability of each environment element in the prior environment information at the next moment, and completing environment information updating. According to the invention, a basis can be provided for an operation process, and prior environment information can be updated.

Description

Technical Field

[0001] The invention relates to an environmental information updating method, device and storage medium, belonging to the technical field of engineering machinery. Background Art

[0002] Unlike unmanned passenger cars, the digital construction of construction machinery has the problem of mutual influence between the construction environment and the equipment operation process. In this case, the environmental factors that hinder the construction process come not only from dynamic targets that intrude into the equipment operation range, but also from the environmental factors that "grow" from the environment itself under the influence of the equipment operation.

[0003] Faced with the above situation, engineering machinery and equipment deployed with a digital construction system need to detect, describe, and manage changes in environmental states, and build a mutual construction mechanism between prior knowledge and a posteriori knowledge of environmental information.

[0004] The invention patent with publication number CN115223118A provides a method for judging the confidence of a high-precision map, including obtaining the real road traffic information of the current location after detecting the use of the high-precision map; based on the real road traffic information, limiting the total elements to obtain the determination elements that need to be judged for confidence; based on the determination elements, obtaining the map traffic information of the current location on the high-precision map, and comparing the map traffic information with the determination elements to obtain the confidence of the determination elements; according to the confidence of the determination elements, evaluating the confidence of the high-precision map. The present invention evaluates the timeliness and reliability of high-precision maps and improves the accuracy of the use of high-precision map information by autonomous driving algorithms.

[0005] This invention patent only considers the guiding role of prior environmental factors on device behavior, and adds confidence attributes and confidence judgment functions on the basis of the guiding role, but does not consider the possible adverse effects of device behavior on environmental factors. Summary of the invention

[0006] The purpose of the present invention is to overcome the deficiencies in the prior art and provide an environmental information updating method, device and storage medium, which can provide a basis for the operation process and can update the prior environmental information. To achieve the above purpose, the present invention is implemented by adopting the following technical solutions:

[0007] In a first aspect, the present invention provides an environment information updating method, comprising:

[0008] Obtain the sensor data of each operating equipment at the current moment and the global probability of each environmental element in the prior environmental information at the current moment;

[0009] Based on the sensor data of the operating equipment at the current moment, each target detected by each operating equipment is obtained, and the local probability of each target is calculated;

[0010] Based on the calculated local probability of each target, the transition probability between the target data of each target and each environmental element is calculated. According to the global probability of each environmental element in the prior environmental information at the current moment and the calculated transition probability, the global probability of each environmental element in the prior environmental information at the next moment is updated to complete the environmental information update.

[0011] In combination with the first aspect, optionally, obtaining each target detected by each operating device based on the sensor data of the operating device at the current moment obtained includes:

[0012] The number of initialized working devices is n, and each device is D(i), where i = 1…n;

[0013] Perform feature extraction on the sensor data of the operating equipment at the current moment to obtain feature data;

[0014] According to the Bayesian theorem, the extracted feature data is tracked to obtain the m detected by each operating device D(i). i targets O(i, j) and corresponding target data and target status data, where i = 1...n, j = 1...m i ; Wherein, the state data of the target includes the expected value of the target state and uncertainty value σ, where i = 1…n, j = 1…m i .

[0015] In combination with the first aspect, optionally, the local probability includes a local existence probability, a local semantic probability and a local behavior probability;

[0016] The calculating of the local probability of each target comprises:

[0017] According to the state data of the target O(i, j), the type of the target is determined to be a point target, a surface target or a volume target; where i = 1...n, j = 1...m i ;

[0018] When the type of target O(i, j) is a point target, calculate the local existence probability of target O(i, j);

[0019] When the type of target O(i, j) is a face target, calculate the local existence probability and local semantic probability of target O(i, j);

[0020] When the type of the target O(i, j) is a body target, the local existence probability, local semantic probability and local behavior probability of the target O(i, j) are calculated.

[0021] In combination with the first aspect, optionally, the calculating the local existence probability, the local semantic probability and the local behavior probability of the target O(i, j) includes:

[0022] Calculate the local existence probability of the target O(i, j), including:

[0023] According to the state data of the target O(i, j), the target state x has the relationship in, is the expected value of the target state, σ is the uncertainty value of the target state;

[0024] The signal-to-noise ratio (SNR) of the target data is a monotonic function of σ, and there exists a relationship SNR = f(σ);

[0025] Based on the local existence probability of the target O(i, j) The relationship between the signal-to-noise ratio (SNR) of the target data Calculate the local existence probability of the target O(i, j) Where g(SNR) is the local existence probability calculation of the signal-to-noise ratio (SNR) of the target data;

[0026] The local semantic probability of the target O(i, j) is calculated using a deep learning network

[0027] Calculate the local behavior probability of the target O(i, j), including:

[0028] According to the local semantic probability of the target O(i, j) Determine the semantic information of the target O(i, j) Among them, X is the semantic set corresponding to the digital construction scene;

[0029] Initialize the semantic set that affects the construction task in the digital construction scene as Calculate the prior probability f1(x) that the semantic information of the target O(i, j) affects the execution of the construction task, using the following formula:

[0030]

[0031] Initialize the construction task T(i) corresponding to the device D(i), and calculate the correlation between the target O(i, j) and the construction task T(i)

[0032] Based on the relative pose vector between the target O(i, j) and the device D(i) The local existence probability of the target O(i, j) The correlation degree d(O(i, j)) between the target O(i, j) and the construction task T(i) is used to calculate the posterior probability that the target O(i, j) affects the execution of the construction task T(i) Among them, F() is used to calculate the posterior probability;

[0033] Calculate the probability that the target O(i, j) with semantic information X′ affects the construction task T(i)

[0034] Local semantic probability based on target O(i, j) and probability Calculate the local behavior probability of target O(i, j)

[0035] In combination with the first aspect, optionally, the calculating of the transition probability between the target data of each target and each environmental element based on the calculated local probability of each target, and updating the global probability of each environmental element in the prior environmental information at the next moment according to the global probability of each environmental element in the acquired prior environmental information at the current moment and the calculated transition probability, includes:

[0036] Calculate the state transition probability between the target data of each target and each environmental element According to the global existence probability of each environmental element in the current moment prior environmental information obtained And the calculated state transition probability Update the global existence probability of each environmental element in the prior environmental information at the next moment

[0037] Calculate the semantic transition probability P(A|A0) between the target data of each target and each environmental element. According to the global semantic probability P(A0) of each environmental element in the prior environmental information at the current moment and the calculated semantic transition probability P(A|A0), update the global semantic probability P(A) of each environmental element in the prior environmental information at the next moment.

[0038] Calculate the behavior transition probability P(ξ|ξ0) between the target data of each target and each environmental element. According to the global behavior probability P(ξ0) of each environmental element in the prior environmental information at the current moment and the calculated behavior transition probability P(ξ|ξ0), update the global behavior probability P(ξ) of each environmental element in the prior environmental information at the next moment.

[0039] In combination with the first aspect, optionally, the state transition probability between the target data of each target and each environmental element is calculated According to the global existence probability of each environmental element in the current moment prior environmental information obtained And the calculated state transition probability Update the global existence probability of each environmental element in the prior environmental information at the next moment include:

[0040] Calculate the state correlation between the target data O(i, j) from the device D(i) and each environmental element E(k) Based on the state association gate Γ(k), the state data association between O(i, j) and E(k) is completed to obtain D(i, j);

[0041] Calculate the state transition probability between the target data O(i, j) from the device D(i) and each environmental element E(k)

[0042] According to DS evidence theory, the state transition probability Fusion is performed to obtain the state transition probability between the target data from the device D(i) and each environmental element E(k) in, For fusion computing;

[0043] According to DS evidence theory, the state transition probability obtained from different devices D(i) Fusion is performed to obtain the state transition probability between each target from each device and each environmental element at the next moment

[0044] According to the global existence probability of each environmental element in the current moment prior environmental information obtained And the state transition probability between each target from each device and each environmental element at the next moment Update the global existence probability of each environmental element E(k) in the prior environmental information at the next moment

[0045] In combination with the first aspect, optionally, the calculating of the semantic transition probability P(A|A0) between the target data of each target and each environmental element, based on the acquired global semantic probability P(A0) of each environmental element in the prior environmental information at the current moment and the calculated semantic transition probability P(A|A0), updates the global semantic probability P(A) of each environmental element in the prior environmental information at the next moment, including:

[0046] Calculate the semantic association between the target data O(i, j) from the device D(i) and each environmental element E(k) Based on the semantic association gate Λ(k), the semantic association between O(i, j) and E(k) is completed, and we get Among them, Y(k) is the semantic set of environmental elements, P(d) is the representative element of semantic probability, is the local semantic probability of target O(i, j);

[0047] Calculate the semantic transition probability between the target data O(i, j) from the device D(i) and each environmental element E(k)

[0048] According to DS evidence theory, the semantic transition probability P ij (A|A0) is fused to obtain the semantic transfer probability between the target detection result from device D(i) and each environmental element in, For fusion computing;

[0049] According to the DS evidence theory, the semantic transition probability P obtained from different devices D(i) i (A|A0) is integrated to obtain the semantic transfer probability of each target and each environmental element from each device at the next moment

[0050] According to the global semantic probability P(A0) of each environmental element in the prior environmental information at the current moment, and the semantic transfer probability P(A|A0) between each target from each device and each environmental element at the next moment, the global semantic probability P(A)=P(A0)P(A|A0) of each environmental element E(k) in the prior environmental information at the next moment is updated.

[0051] In combination with the first aspect, optionally, the calculating of the behavior transition probability P(ξ|ξ0) between the target data of each target and each environmental element, based on the global behavior probability P(ξ0) of each environmental element in the prior environmental information at the current moment obtained and the calculated behavior transition probability P(ξ|ξ0), updates the global behavior probability P(ξ) of each environmental element in the prior environmental information at the next moment, including:

[0052] Calculate the behavioral correlation between the construction task T(i) corresponding to each different device D(i) and each environmental factor E(k)

[0053] according to Calculate the probability of the behavior association between device D(i) and each environmental element E(k) at the next moment Among them, F() is the behavior association probability calculation;

[0054] Behavior-related gate The behavior association probability P obtained by T(i, k), for a specific prior environmental factor E(k), the degree of association is graded for different devices D(i);

[0055] At the behavioral relevance level L t In the calculation, the correlation between the target data O(i, j) from the device D(i) and each environmental factor E(k) is calculated. Based on the state association gate Γ(k), complete the data association between O(i, j) and E(k) to obtain D(i, j);

[0056] At the behavioral relevance level L t According to the DS evidence theory, the behavior transition probability between the target data O(i, j) from the device D(i) and each environmental factor E(k) is calculated. in, For fusion computing;

[0057] At the behavioral relevance level L t According to the DS evidence theory, at the construction task association level L t In the example, for different devices D(i), the behavior transition probability P ti (ξ|ξ0) is integrated to obtain the behavior transfer probability of each target and each environmental element from each device at the next moment

[0058] According to DS evidence theory, different behavior association levels L t For all devices D(i) in the t (ξ|ξ0) is integrated to obtain the behavior transfer probability of each target and each environmental element from each device at the next moment

[0059] According to the global behavior probability P(ξ0) of each environmental element in the prior environmental information at the current moment, and the behavior transfer probability P(ξ|ξ0) between each target from each device and each environmental element at the next moment, the global behavior probability P(ξ)=P(ξ0)P(ξ|ξ0) of each environmental element E(k) in the prior environmental information at the next moment is updated.

[0060] In a second aspect, the present invention provides an environment information updating device, comprising:

[0061] Acquisition module: used to obtain the sensor data of each operating equipment at the current moment and the global probability of each environmental element in the prior environmental information at the current moment;

[0062] The first calculation module is used to obtain each target detected by each operating device based on the sensor data of the operating device at the current moment, and calculate the local probability of each target;

[0063] Calculation and update module: used to calculate the transfer probability between the target data of each target and each environmental element based on the calculated local probability of each target, and update the global probability of each environmental element in the prior environmental information at the next moment according to the global probability of each environmental element in the current moment obtained and the calculated transfer probability, so as to complete the update of environmental information.

[0064] In a third aspect, the present invention provides a computer-readable storage medium having a computer program / instruction stored thereon, which, when executed by a processor, implements the steps of the environmental information updating method described in the first aspect.

[0065] Compared with the prior art, the environmental information updating method, device and storage medium provided by the embodiments of the present invention have the following beneficial effects:

[0066] The present invention obtains the sensor data of each operating device at the current moment and the global probability of each environmental element in the prior environmental information at the current moment; the present invention can provide a basis for the operating process based on the sensor data and the prior environmental information to ensure normal operation;

[0067] The present invention obtains each target detected by each operating device based on the sensor data of the operating device at the current moment, and calculates the local probability of each target; based on the calculated local probability of each target, the transition probability between the target data of each target and each environmental element is calculated, and the global probability of each environmental element in the prior environmental information at the next moment is updated according to the global probability of each environmental element in the prior environmental information at the current moment obtained and the calculated transition probability; the present invention can update the prior environmental information;

[0068] The present invention collects sensor data and prior environmental information in real time, and updates the prior environmental information in real time, taking into account the relationship between the environment and the operating equipment, and can improve the efficiency and stability of the operating process and enhance the safety of the system. BRIEF DESCRIPTION OF THE DRAWINGS

[0069] Figure 1 is a flow chart of a method for updating environmental information in Embodiment 1 of the present invention;

[0070] Figure 2 is a flow chart of calculating the local existence probability in an environmental information updating method in Embodiment 1 of the present invention;

[0071] Figure 3is a flow chart of calculating the probability of a local behavior in a method for updating environmental information in Embodiment 1 of the present invention;

[0072] Figure 4 It is a flow chart of calculating state transition probability and updating global existence probability in an environmental information updating method in embodiment 1 of the present invention;

[0073] Figure 5 It is a flow chart of calculating semantic transition probability and updating global semantic probability in an environmental information updating method in embodiment 1 of the present invention;

[0074] Figure 6 It is a flow chart of calculating behavior transition probability and updating global behavior probability in an environmental information updating method in embodiment 1 of the present invention;

[0075] Figure 7 This is a flowchart of the prediction process of the environment status layer where the target data is located in an environment information updating method in Example 2 of the present invention.

[0076] Figure 8 It is a flow chart of the updating process of the environment status layer where the environment elements are located in an environment information updating method in Embodiment 2 of the present invention;

[0077] Fig. 9 It is a state transition diagram of environmental state changes defined by an environmental state layer in an environmental information updating method in embodiment 2 of the present invention. DETAILED DESCRIPTION

[0078] The present invention will be further described below in conjunction with the accompanying drawings. The following embodiments are only used to more clearly illustrate the technical solution of the present invention, and cannot be used to limit the protection scope of the present invention.

[0079] Embodiment 1:

[0080] like Figure 1 As shown, this embodiment provides a method for updating environment information, including:

[0081] Obtain the sensor data of each operating equipment at the current moment and the global probability of each environmental element in the prior environmental information at the current moment;

[0082] Based on the sensor data of the operating equipment at the current moment, each target detected by each operating equipment is obtained, and the local probability of each target is calculated;

[0083] Based on the calculated local probability of each target, the transition probability between the target data of each target and each environmental element is calculated. According to the global probability of each environmental element in the prior environmental information at the current moment and the calculated transition probability, the global probability of each environmental element in the prior environmental information at the next moment is updated to complete the environmental information update.

[0084] The specific steps include:

[0085] Step 1: Obtain the sensor data of each operating equipment at the current moment and the global probability of each environmental element in the prior environmental information at the current moment.

[0086] Step 2: Based on the sensor data of the operating equipment at the current moment, obtain each target detected by each operating equipment and calculate the local probability of each target.

[0087] Step 2.1: Based on the sensor data of the operating equipment at the current moment, obtain each target detected by each operating equipment.

[0088] Step 2.1.1: Initialize the number of operating devices to n, each device is D(i), where i = 1...n.

[0089] Step 2.1.2: Extract features from the sensor data of the operating equipment at the current moment to obtain feature data.

[0090] Step 2.1.3: According to Bayesian theorem, the extracted feature data is tracked to obtain the m detected by each operating device D(i). i targets O(i, j) and corresponding target data and target status data, where i = 1...n, j = 1...m i ; Wherein, the state data of the target includes the expected value of the target state and uncertainty value σ, where i = 1…n, j = 1…m i .

[0091] Step 2.2: Calculate the local probability of each target.

[0092] For the local probability of the target, it specifically corresponds to the following three problems.

[0093] Question 1: Is the target stable for the device itself? The answer to this question can generate the local existence probability of the target. describe.

[0094] Question 2: For the device itself, what type of target is it? The answer to this question can generate the local semantic probability of the target. describe.

[0095] Question 3: For the equipment itself, the relationship between the target and the construction task. The answer to this question can generate the local behavior probability of the target. describe.

[0096] Therefore, the local probability includes the local existence probability, the local semantic probability and the local behavior probability.

[0097] According to the state data of the target O(i, j), the type of the target is determined to be a point target, a surface target or a volume target; where i = 1...n, j = 1...m i .

[0098] When the type of the target O(i, j) is a point target, the local existence probability of the target O(i, j) is calculated.

[0099] When the type of the target O(i, j) is a face target, the local existence probability and local semantic probability of the target O(i, j) are calculated.

[0100] When the type of the target O(i, j) is a body target, the local existence probability, local semantic probability and local behavior probability of the target O(i, j) are calculated.

[0101] For a specific sensor detection result, it must contain two parts: true value and noise. The local existence probability of the target It can be defined as a function of the signal-to-noise ratio (SNR) of the measurement result. Figure 2 As shown, the local existence probability of the target O(i, j) is calculated, including:

[0102] According to the state data of the target O(i, j), the target state x has the relationship in, is the expected value of the target state, σ is the uncertainty value of the target state;

[0103] The signal-to-noise ratio (SNR) of the target data is a monotonic function of σ, and there exists a relationship SNR = f(σ);

[0104] Based on the local existence probability of the target O(i, j) The relationship between the signal-to-noise ratio (SNR) of the target data Calculate the local existence probability of the target O(i, j) Where g(SNR) is the local existence probability calculation of the signal-to-noise ratio (SNR) of the target data;

[0105] If the global probability of each environmental element in the prior environmental information at the current moment does not exist, the global probability of each environmental element in the prior environmental information at the current moment is calculated.

[0106] For some sensors, the type of the target can be determined by the optical features or color features from the raw data of the sensor, and the probability that the target belongs to a specific type is given, which is called local semantic probability. Calculate the local semantic probability of the target O(i, j), including: using the deep learning network to calculate the local semantic probability of the target O(i, j)

[0107] Local behavior probability for the target Break it down into the following questions:

[0108] First, the probability that the target belongs to a specific semantics is calculated using the local semantic probability of the target to describe;

[0109] Second, the probability that a specific semantic target affects the construction task The calculation is done according to the following relationship: Among them, f1(x) is the prior probability that a specific semantic x affects the execution of the construction task, the semantic set is X, and the semantic set that will affect the construction task is f2(O) is the posterior probability that target O affects the execution of construction task T.

[0110] like Figure 3 As shown, the local behavior probability of the target O(i, j) is calculated, including:

[0111] According to the local semantic probability of the target O(i, j) Determine the semantic information of the target O(i, j) Among them, X is the semantic set corresponding to the digital construction scene;

[0112] Initialize the semantic set that affects the construction task in the digital construction scene as Calculate the prior probability f1(x) that the semantic information of the target O(i, j) affects the execution of the construction task, using the following formula:

[0113]

[0114] Initialize the construction task T(i) corresponding to the device D(i), and calculate the correlation between the target O(i, j) and the construction task T(i)

[0115] Based on the relative pose vector between the target O(i, j) and the device D(i) The local existence probability of the target O(i, j) The correlation degree d(O(i, j)) between the target O(i, j) and the construction task T(i) is used to calculate the posterior probability that the target O(i, j) affects the execution of the construction task T(i) Among them, F() is used to calculate the posterior probability;

[0116] Calculate the probability that the target O(i, j) with semantic information X′ affects the construction task T(i)

[0117] Local semantic probability based on target O(i, j) and probability Calculate the local behavior probability of target O(i, j)

[0118] Step 3: Based on the calculated local probabilities of each target, calculate the transition probabilities between the target data of each target and each environmental element. According to the global probabilities of each environmental element in the prior environmental information at the current moment and the calculated transition probabilities, update the global probabilities of each environmental element in the prior environmental information at the next moment to complete the environmental information update.

[0119] Step 3.1: When the global probability of each environmental element in the prior environmental information at the current moment does not exist, calculate the global probability of each environmental element. At time t, the construction task corresponding to different operating equipment D(i) is T(i), where i=1…n.

[0120] The global probability of environmental elements corresponds to the following three questions.

[0121] Question 1: Globally, the life cycle of environmental factors includes the time of their occurrence and the time of their extinction. The answer to this question can generate the probability of existence of environmental factors. describe.

[0122] Question 2: Globally, what type of environmental elements are they? The answer to this question can generate a semantic probability description P(A) of the environmental elements.

[0123] Question 3: Globally, the relationship between environmental factors and construction tasks. The answer to this question can produce a description of the behavior probability P(ξ) of environmental factors.

[0124] The global probability of each environmental factor is calculated based on this, and the specific steps include:

[0125] Step 3.1.1: Calculate the state correlation between the target data of each operating equipment and the prior environmental elements, and complete the state data correlation between the target data of each operating equipment and the prior environmental elements.

[0126] The state correlation D(i, j, k) between the target detection result O(i, j) from the device D(i) and the prior environmental element E(k) is as follows:

[0127]

[0128] According to the required correlation degree D(i, j, k), the targets falling into the state correlation gate Γ(k) are screened, and a total of j′ targets fall into the state correlation gate T(k).

[0129] By filtering through the state association gate, the data association between the target data O(i, j) and the prior environmental elements E(k) is completed, thereby reducing the dimension of the association degree and turning D(i, j, k) into D(i, j).

[0130] Step 3.1.2: Calculate the semantic association between the target data of each operating device and the prior environmental elements, and complete the semantic association between the target data of each operating device and the prior environmental elements.

[0131] The semantic set corresponding to the target data O(i, j) from device D(i) is X(i, j), and the semantic set corresponding to the prior environmental factor E(k) is Y(k). Then the semantic association between O(i, j) and E(k) is as follows:

[0132]

[0133] According to the semantic association gate Λ(k), the semantic association is completed by filtering the semantic information that falls into it, and the dimension of the semantic association is reduced, so that becomes

[0134] Step 3.1.3: Calculate the behavioral correlation between the construction task T(i) corresponding to each different device D(i) and the prior environmental factor E(k) As shown in the following formula,

[0135]

[0136] Then according to The probability of association between the computing device D(i) and the prior environmental factor E(k) at time t0+Δt conforms to the following relationship:

[0137]

[0138] Then, according to the behavior association gate The behavior association probability P is obtained T (i, k), for a specific priori environmental factor E(k), the association degree of different devices D(i) is graded, which can be divided into τ behavior association levels, denoted by L t , (where t=1…τ).

[0139] Step 3.2: Calculate the state transition probability between the target data of each target and each environmental element According to the global existence probability of each environmental element in the current moment prior environmental information obtained And the calculated state transition probability Update the global existence probability of each environmental element in the prior environmental information at the next moment

[0140] like Figure 4 As shown, including:

[0141] Step 3.2.1: Calculate the state correlation between the target data O(i, j) from the device D(i) and each environmental element E(k) Based on the state association gate Γ(k), the state data association between O(i, j) and E(k) is completed to obtain D(i, j).

[0142] Step 3.2.2: Calculate the state transition probability between the target data O(i, j) from the device D(i) and each environmental element E(k)

[0143] Step 3.2.3: According to DS evidence theory, the state transition probability Fusion is performed to obtain the state transition probability between the target data from the device D(i) and each environmental element E(k) in, For fusion calculation.

[0144] Step 3.2.4: According to DS evidence theory, the state transition probability obtained from different devices D(i) Fusion is performed to obtain the state transition probability between each target from each device and each environmental element at the next moment

[0145] Step 3.2.5: Based on the global existence probability of each environmental element in the current prior environmental information obtained And the state transition probability between each target from each device and each environmental element at the next moment Update the global existence probability of each environmental element E(k) in the prior environmental information at the next moment

[0146] Step 3.3: Calculate the semantic transfer probability P(A|A0) between the target data of each target and each environmental element. According to the global semantic probability P(A0) of each environmental element in the prior environmental information at the current moment and the calculated semantic transfer probability P(A|A0), update the global semantic probability P(A) of each environmental element in the prior environmental information at the next moment.

[0147] like Figure 5 As shown, including:

[0148] Step 3.3.1: Calculate the semantic association between the target data O(i, j) from the device D(i) and each environmental element E(k) Based on the semantic association gate Λ(k), the semantic association between O(i, j) and E(k) is completed, and we get Among them, Y(k) is the semantic set of environmental elements, P(d) is the representative element of semantic probability, is the local semantic probability of target O(i, j).

[0149] Step 3.3.2: Calculate the semantic transition probability between the target data O(i, j) from the device D(i) and each environmental element E(k)

[0150] Step 3.3.3: According to DS evidence theory, the obtained semantic transition probability P ij (A|A0) is fused to obtain the semantic transfer probability between the target detection result from device D(i) and each environmental element in, For fusion calculation.

[0151] Step 3.3.4: According to DS evidence theory, the semantic transition probability P obtained from different devices D(i) i (A|A0) is integrated to obtain the semantic transfer probability of each target and each environmental element from each device at the next moment

[0152] Step 3.3.5: Based on the global semantic probability P(A0) of each environmental element in the prior environmental information at the current moment, and the semantic transfer probability P(A|A0) between each target from each device and each environmental element at the next moment, update the global semantic probability P(A)=P(A0)P(A|A0) of each environmental element E(k) in the prior environmental information at the next moment.

[0153] Step 3.4: Calculate the behavior transition probability P(ξ|ξ0) between the target data of each target and each environmental element. According to the global behavior probability P(ξ0) of each environmental element in the prior environmental information at the current moment and the calculated behavior transition probability P(ξ|ξ0), update the global behavior probability P(ξ) of each environmental element in the prior environmental information at the next moment.

[0154] like Figure 6 As shown, including:

[0155] Step 3.4.1: Calculate the behavioral correlation between the construction task T(i) corresponding to each different device D(i) and each environmental factor E(k)

[0156] Step 3.4.2: According to Calculate the probability of the behavior association between device D(i) and each environmental element E(k) at the next moment Among them, F() is the behavior association probability calculation.

[0157] Step 3.4.3: Behavior-based association gate The behavior association probability P obtained by T (i, k), for a specific priori environmental factor E(k), the degree of association is graded for different devices D(i).

[0158] Step 3.4.4: At the behavior association level L t In the calculation, the correlation between the target data O(i, j) from the device D(i) and each environmental factor E(k) is calculated. Based on the state association gate Γ(k), the data association between O(i, j) and E(k) is completed to obtain D(i, j).

[0159] Step 3.4.5: At the behavior association level L t According to the DS evidence theory, the behavior transition probability between the target data O(i, j) from the device D(i) and each environmental factor E(k) is calculated. in, For fusion calculation.

[0160] Step 3.4.6: At the behavior association level L t According to the DS evidence theory, at the construction task association level L t In the example, for different devices D(i), the behavior transition probability P ti (ξ|ξ0) is integrated to obtain the behavior transfer probability of each target and each environmental element from each device at the next moment

[0161] Step 3.4.7: According to DS evidence theory, different behavior association levels L t For all devices D(i) in the t (ξ|ξ0) is integrated to obtain the behavior transfer probability of each target and each environmental element from each device at the next moment

[0162] Step 3.4.8: Based on the global behavior probability P(ξ0) of each environmental element in the prior environmental information at the current moment, and the behavior transfer probability P(ξ|ξ0) between each target from each device and each environmental element at the next moment, update the global behavior probability P(ξ)=P(ξ0)P(ξ|ξ0) of each environmental element E(k) in the prior environmental information at the next moment.

[0163] The sensor data and prior environmental information collected in real time in this embodiment provide a basis for the normal progress of the operation process; the prior environmental information can be updated, which can improve the efficiency and stability of the operation process and enhance the security of the system.

[0164] Embodiment 2:

[0165] Based on the inventive concept of Example 1, this embodiment provides a method of dividing the description of the environmental status into three environmental status layers: foreground, middle ground and background, according to the hindering effect of environmental factors on the equipment operation process.

[0166] Prospects include environmental factors that will inevitably hinder the operation of the equipment during operation. Prospect targets include two types:

[0167] First, the dynamic target that breaks into the equipment's operating range is the foreground dynamic target;

[0168] Secondly, during the equipment operation process, the static targets generated by the operating environment under the influence of the operating process are foreground static targets.

[0169] For prospective targets, there are four sources:

[0170] First, it is generated randomly during the operation;

[0171] Second, it is transformed from the mid-range target;

[0172] Third, it is directly transformed from the background target;

[0173] Fourth, its own potential transformation.

[0174] The mid-view includes potential environmental factors that may hinder the operation process during the equipment operation. The mid-view target refers to the target that already exists in the operation environment and will not hinder the operation process at present, but as the operation process progresses, its hindering effect will gradually become apparent.

[0175] For mid-range targets, there are four sources:

[0176] First, it is generated randomly during the operation;

[0177] Second, it is derived from the degradation of foreground targets;

[0178] Third, it is transformed from background targets;

[0179] Fourth, its own potential changes.

[0180] Background includes environmental factors that will not hinder the operation of the equipment during operation. The background target itself will not constitute an obstacle.

[0181] For background targets, there are four sources:

[0182] First, the innate existence independent of the operational process;

[0183] Second, it is directly degraded from the foreground target;

[0184] Third, it is derived from the degradation of mid-ground targets;

[0185] Fourth, its own potential changes.

[0186] Based on the above environmental state division method, the environmental state changes come from four aspects:

[0187] First, the random occurrence, transformation and disappearance of foreground dynamic targets;

[0188] Second, the random occurrence, transformation, and disappearance of foreground static objects;

[0189] Third, the emergence, change and disappearance of mid-ground targets;

[0190] Fourth, the potential changes and disappearance of background targets.

[0191] This embodiment provides an environment information updating method based on the environment layer, including:

[0192] Step 1: Obtain the sensor data of each operating equipment at the current moment and the global probability of each environmental element in the prior environmental information at the current moment.

[0193] Step 2: Initialize the environmental state layer where each environmental element in the prior environmental information at the current moment is located.

[0194] Step 3: Based on the sensor data of the operating equipment at the current moment, obtain each target detected by each operating equipment and calculate the local probability of each target.

[0195] Step 4: Determine the type of the target. When the target type is a body target, determine its local semantic probability. and local behavior probability The environmental status layer to which the prediction target belongs.

[0196] Specific steps are as follows Figure 7 As shown:

[0197] Step 4.1: Based on the local semantic probability of the target O(i, j) Determine the semantic information of the target O(i, j), Among them, the semantic set corresponding to the digital construction scene is X.

[0198] Step 4.2: Initialize the semantic set that affects the construction task in the digital construction scene as The prior probability that the semantic information of the target O(i, j) affects the execution of the construction task is calculated by the following relation:

[0199]

[0200] Step 4.3: The local behavior probability based on the target O(i, j) is

[0201] If f1(x)=1, and Greater than a given threshold Right now Then the target O(i, j) belongs to the foreground target,

[0202] If f1(x)=1, and Greater than a given threshold but not exceeding a given threshold Right now Then the target O(i, j) belongs to the mid-view target,

[0203] If f1(x) = 0, or Does not exceed a given threshold Right now Then the target O(i, j) belongs to the background target.

[0204] Step 5: Based on the calculated local probabilities of each target, calculate the transition probabilities between the target data of each target and each environmental element. According to the global probabilities of each environmental element in the prior environmental information at the current moment and the calculated transition probabilities, update the global probabilities of each environmental element in the prior environmental information at the next moment to complete the environmental information update.

[0205] The local probability of the target data is specifically fused with the global probability of the prior environmental elements. Specifically: the local probability of the foreground local target is used to fuse the global probability of the foreground prior environmental elements; the local probability of the midground local target is used to fuse the global probability of the midground prior environmental elements; the local probability of the background local target is used to fuse the global probability of the background prior environmental elements.

[0206] Step 6: Update the environmental state layer where the prior environmental element is located at the next moment according to the global semantic probability and global behavior probability of the prior environmental element at the next moment.

[0207] Specific steps are as follows Figure 8 As shown:

[0208] Step 6.1: According to the local semantic probability P(A) of the target O(i, j), determine the semantic information of the prior environmental element E(k), X′={x0∈X|P(x0)≥P(A)), where the semantic set corresponding to the digital construction scene is X.

[0209] Step 6.2: The semantic set that affects the construction task in the digital construction scene is The prior probability that the semantic information of the target O(i, j) affects the execution of the construction task can be calculated by the following relationship:

[0210]

[0211] Step 6.3: The global behavior probability of the prior environmental element E(k) is P(ξ),

[0212] like And P(ξ) is greater than the given threshold P1, that is, P(ξ)>P1, then the prior environmental element E(k) belongs to the foreground target;

[0213] like And P(ξ) is greater than the given threshold P00, but does not exceed the given threshold P1, that is, P0<P(ξ)≤P1, then the prior environmental element E(k) belongs to the mid-ground target;

[0214] like Or P(ξ) does not exceed the given threshold P0, that is, P(ξ)≤P0, then the prior environmental element E(k) belongs to the background target.

[0215] After completing the attribute fusion, the global semantic probability and global behavior probability of the updated environmental element are used to determine again which environmental state layer it belongs to, and the environmental state layer where the prior environmental element is located at the next moment is updated.

[0216] like Fig. 9 As shown, in this embodiment, there are 13 possible forms of environmental state changes based on the environmental state layer:

[0217] (1) The process in which the foreground target represented by sequence number 1 in the figure occurs randomly;

[0218] (2) The state transition process of the foreground object represented by number 2 in the figure within this layer;

[0219] (3) The process in which the foreground object represented by number 3 disappears;

[0220] (4) The process of transforming the mid-ground target represented by number 4 into the foreground target;

[0221] (5) The process in which the foreground target represented by number 5 in the figure degenerates into a mid-ground target;

[0222] (6) The process in which the background target represented by number 6 in the figure is directly transformed into the foreground target;

[0223] (7) The process in which the foreground target represented by number 7 in the figure directly degenerates into a background target;

[0224] (8) The random occurrence process of the mid-ground target represented by sequence number 8 in the figure;

[0225] (9) The state transition process of the mid-ground target represented by number 9 in the figure within this layer;

[0226] (10) The process in which the mid-ground target represented by number 10 disappears;

[0227] (11) The process of transforming the background target represented by number 11 into the mid-ground target;

[0228] (12) The process in which the mid-ground target represented by sequence number 12 in the figure degenerates into a background target;

[0229] (13) The background update process represented by sequence number 13 in the figure.

[0230] This embodiment takes into account the relationship between the environment and the operating equipment, and can improve the efficiency and stability of the operating process while enhancing the safety of the system.

[0231] Embodiment 3:

[0232] This embodiment provides an environment information updating device, including:

[0233] Acquisition module: used to obtain the sensor data of each operating equipment at the current moment and the global probability of each environmental element in the prior environmental information at the current moment;

[0234] The first calculation module is used to obtain each target detected by each operating device based on the sensor data of the operating device at the current moment, and calculate the local probability of each target;

[0235] Calculation and update module: used to calculate the transfer probability between the target data of each target and each environmental element based on the calculated local probability of each target, and update the global probability of each environmental element in the prior environmental information at the next moment according to the global probability of each environmental element in the current moment obtained and the calculated transfer probability, so as to complete the update of environmental information.

[0236] Embodiment 4:

[0237] This embodiment provides a computer-readable storage medium on which a computer program / instruction is stored. When the computer program / instruction is executed by a processor, the steps of the environmental information updating method described in Embodiment 1 or Embodiment 2 are implemented.

[0238] Those skilled in the art will appreciate that embodiments of the present invention may be provided as methods, systems, or computer program products. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0239] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0240] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0241] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0242] The embodiments of the present invention are described above in conjunction with the accompanying drawings, but the present invention is not limited to the above-mentioned specific implementation methods. The above-mentioned specific implementation methods are merely illustrative and not restrictive. Under the enlightenment of the present invention, ordinary technicians in this field can also make many forms without departing from the scope of protection of the purpose of the present invention and the claims, which all fall within the protection of the present invention.

Claims

1. A method for updating environmental information, characterized in that: include: Obtain the sensor data of each operating equipment at the current moment and the global probability of each environmental element in the prior environmental information at the current moment; Based on the sensor data of the operating equipment at the current moment, each target detected by each operating equipment is obtained, and the local probability of each target is calculated; Based on the calculated local probability of each target, the transition probability between the target data of each target and each environmental element is calculated. According to the global probability of each environmental element in the prior environmental information at the current moment and the calculated transition probability, the global probability of each environmental element in the prior environmental information at the next moment is updated to complete the environmental information update.

2. The method for updating environmental information according to claim 1, characterized in that: The method of obtaining each target detected by each operating device based on the sensor data of the operating device at the current moment includes: The number of initialized working devices is n, and each device is D(i), where i = 1…n; Perform feature extraction on the sensor data of the operating equipment at the current moment to obtain feature data; According to the Bayesian theorem, the extracted feature data is tracked to obtain the m detected by each operating device D(i). i targets O(i, j) and corresponding target data and target status data, where i = 1...n, j = 1...m i ; Wherein, the state data of the target includes the expected value of the target state and uncertainty value σ, where i = 1…n, j = 1…m i .

3. The method for updating environmental information according to claim 1, characterized in that: The local probability includes local existence probability, local semantic probability and local behavior probability; The calculating of the local probability of each target comprises: According to the state data of the target O(i, j), the type of the target is determined to be a point target, a surface target or a volume target; where i = 1...n, j = 1...m i ; When the type of target O(i, j) is a point target, calculate the local existence probability of target O(i, j); When the type of target O(i, j) is a face target, calculate the local existence probability and local semantic probability of target O(i, j); When the type of the target O(i, j) is a body target, the local existence probability, local semantic probability and local behavior probability of the target O(i, j) are calculated.

4. The method for updating environmental information according to claim 3, characterized in that: The calculation of the local existence probability, local semantic probability and local behavior probability of the target O(i, j) includes: Calculate the local existence probability of the target O(i, j), including: According to the state data of the target O(i, j), the target state x has the relationship in, is the expected value of the target state, σ is the uncertainty value of the target state; The signal-to-noise ratio (SNR) of the target data is a monotonic function of σ, and there exists a relationship SNR = f(σ); Based on the local existence probability of the target O(i, j) The relationship between the signal-to-noise ratio (SNR) of the target data Calculate the local existence probability of the target O(i, j) Where g(SNR) is the local existence probability calculation of the signal-to-noise ratio (SNR) of the target data; The local semantic probability of the target O(i, j) is calculated using a deep learning network Calculate the local behavior probability of the target O(i, j), including: According to the local semantic probability of the target O(i, j) Determine the semantic information of the target O(i, j) Among them, X is the semantic set corresponding to the digital construction scene; Initialize the semantic set that affects the construction task in the digital construction scene as Calculate the prior probability f1(x) that the semantic information of the target O(i, j) affects the execution of the construction task, using the following formula: Initialize the construction task T(i) corresponding to the device D(i), and calculate the correlation between the target O(i, j) and the construction task T(i) Based on the relative pose vector between the target O(i, j) and the device D(i) The local existence probability of the target O(i, j) The correlation degree d(O(i, j)) between the target O(i, j) and the construction task T(i) is used to calculate the posterior probability that the target O(i, j) affects the execution of the construction task T(i) Among them, F() is used to calculate the posterior probability; Calculate the probability that the target O(i, j) with semantic information X′ affects the construction task T(i) Local semantic probability based on target O(i, j) and probability Calculate the local behavior probability of target O(i, j) 5. The method for updating environmental information according to claim 1, characterized in that: The method calculates the transition probability between the target data of each target and each environmental element based on the calculated local probability of each target, and updates the global probability of each environmental element in the prior environmental information at the next moment according to the global probability of each environmental element in the acquired prior environmental information at the current moment and the calculated transition probability, including: Calculate the state transition probability between the target data of each target and each environmental element According to the global existence probability of each environmental element in the current moment prior environmental information obtained And the calculated state transition probability Update the global existence probability of each environmental element in the prior environmental information at the next moment Calculate the semantic transition probability P(A|A0) between the target data of each target and each environmental element. According to the global semantic probability P(A0) of each environmental element in the prior environmental information at the current moment and the calculated semantic transition probability P(A|A0), update the global semantic probability P(A) of each environmental element in the prior environmental information at the next moment. Calculate the behavior transition probability P(ξ|ξ0) between the target data of each target and each environmental element. According to the global behavior probability P(ξ0) of each environmental element in the prior environmental information at the current moment and the calculated behavior transition probability P(ξ|ξ0), update the global behavior probability P(ξ) of each environmental element in the prior environmental information at the next moment.

6. The method for updating environmental information according to claim 5, characterized in that: The calculation of the state transition probability between the target data of each target and each environmental element According to the global existence probability of each environmental element in the current moment prior environmental information obtained And the calculated state transition probability Update the global existence probability of each environmental element in the prior environmental information at the next moment include: Calculate the state correlation between the target data O(i, j) from the device D(i) and each environmental element E(k) Based on the state association gate Γ(k), the state data association between O(i, j) and E(k) is completed to obtain D(i, j); Calculate the state transition probability between the target data O(i, j) from the device D(i) and each environmental element E(k) According to DS evidence theory, the state transition probability Fusion is performed to obtain the state transition probability between the target data from the device D(i) and each environmental element E(k) in, For fusion computing; According to DS evidence theory, the state transition probability obtained from different devices D(i) Fusion is performed to obtain the state transition probability between each target from each device and each environmental element at the next moment According to the global existence probability of each environmental element in the current moment prior environmental information obtained And the state transition probability between each target from each device and each environmental element at the next moment Update the global existence probability of each environmental element E(k) in the prior environmental information at the next moment 7. The method for updating environmental information according to claim 5, characterized in that: The calculation of the semantic transition probability P(A|A0) between the target data of each target and each environmental element, and updating the global semantic probability P(A0) of each environmental element in the prior environmental information at the next moment according to the global semantic probability P(A0) of each environmental element in the current moment prior environmental information obtained and the calculated semantic transition probability P(A|A0), includes: Calculate the semantic association between the target data O(i, j) from the device D(i) and each environmental element E(k) Based on the semantic association gate Λ(k), the semantic association between O(i, j) and E(k) is completed, and we get Among them, Y(k) is the semantic set of environmental elements, P(d) is the representative element of semantic probability, is the local semantic probability of target O(i, j); Calculate the semantic transition probability between the target data O(i, j) from the device D(i) and each environmental element E(k) According to DS evidence theory, the semantic transition probability P ij (A|A0) is fused to obtain the semantic transfer probability between the target detection result from device D(i) and each environmental element in, For fusion computing; According to the DS evidence theory, the semantic transition probability P obtained from different devices D(i) i (A|A0) is integrated to obtain the semantic transfer probability of each target and each environmental element from each device at the next moment According to the global semantic probability P(A0) of each environmental element in the prior environmental information at the current moment, and the semantic transfer probability P(A|A0) between each target from each device and each environmental element at the next moment, the global semantic probability P(A)=P(A0)P(A|A0) of each environmental element E(k) in the prior environmental information at the next moment is updated.

8. The method for updating environmental information according to claim 5, characterized in that: The calculation of the behavior transition probability P(ξ|ξ0) between the target data of each target and each environmental element, based on the global behavior probability P(ξ0) of each environmental element in the prior environmental information at the current moment obtained and the calculated behavior transition probability P(ξ|ξ0), updates the global behavior probability P(ξ) of each environmental element in the prior environmental information at the next moment, including: Calculate the behavioral correlation between the construction task T(i) corresponding to each different device D(i) and each environmental factor E(k) according to Calculate the probability of the behavior association between device D(i) and each environmental element E(k) at the next moment Among them, F() is the behavior association probability calculation; Behavior-related gate The behavior association probability P obtained by T (i, k), for a specific prior environmental factor E(k), the degree of association is graded for different devices D(i); At the behavioral relevance level L t In the calculation, the correlation between the target data O(i, j) from the device D(i) and each environmental factor E(k) is calculated. Based on the state association gate Γ(k), complete the data association between O(i, j) and E(k) to obtain D(i, j); At the behavioral relevance level L t According to the DS evidence theory, the behavior transition probability between the target data O(i, j) from the device D(i) and each environmental factor E(k) is calculated. in, For fusion computing; At the behavioral relevance level L t According to the DS evidence theory, at the construction task association level L t In the example, for different devices D(i), the behavior transition probability P ti (ξ|ξ0) is integrated to obtain the behavior transfer probability of each target and each environmental element from each device at the next moment According to DS evidence theory, different behavior association levels L t For all devices D(i) in the t (ξ|ξ0) is integrated to obtain the behavior transfer probability of each target and each environmental element from each device at the next moment According to the global behavior probability P(ξ0) of each environmental element in the prior environmental information at the current moment, and the behavior transfer probability P(ξ|ξ0) between each target from each device and each environmental element at the next moment, the global behavior probability P(ξ)=P(ξ0)P(ξ|ξ0) of each environmental element E(k) in the prior environmental information at the next moment is updated.

9. An environmental information updating device, characterized in that: include: Acquisition module: used to obtain the sensor data of each operating equipment at the current moment and the global probability of each environmental element in the prior environmental information at the current moment; The first calculation module is used to obtain each target detected by each operating device based on the sensor data of the operating device at the current moment, and calculate the local probability of each target; Calculation and update module: used to calculate the transfer probability between the target data of each target and each environmental element based on the calculated local probability of each target, and update the global probability of each environmental element in the prior environmental information at the next moment according to the global probability of each environmental element in the current moment obtained and the calculated transfer probability, so as to complete the update of environmental information.

10. A computer-readable storage medium having a computer program / instruction stored thereon, characterized in that: When the computer program / instruction is executed by a processor, the steps of the environmental information updating method described in any one of claims 1-9 are implemented.

Images