Intelligent navigation and photographing robot based on large model optimization

Abstract

The invention relates to the technical field of service robots, in particular to an intelligent guiding and photographing robot based on large model optimization, which comprises a main body, and the main body carries a robot control system. The system comprises a path optimization and dynamic adjustment module, an emotion perception and interaction content adjustment module, a guide content real-time updating module and a photographing time and position selection module. According to the invention, through dynamic data analysis, high efficiency and practicability of path selection are ensured, and the ability to respond to environment change in real time, the fluency and accuracy of the tour guide process are significantly improved, unnecessary delay is avoided, and the interaction content is adjusted by analyzing tourist behavior response data, so that the tour guide efficiency is improved. The robot can more accurately adjust service contents to match moods and interests of tourists, timeliness and correlation of information services are further improved through real-time updating of the guide contents and accurate selection of photographing time, and it is ensured that the tourists obtain valuable visual records and guide information at the best moment.

Description

Technical Field

[0001] The present invention relates to the technical field of service robots, and in particular to an intelligent tour guide and photo-taking robot based on large model optimization. Background Art

[0002] The field of service robotics technology primarily involves the design, construction, and application of robots that can perform human service tasks. These typically include cleaning, delivery, surveillance, nursing, and education. Service robots can be divided into two categories: personal service robots and professional service robots. Personal service robots are often used in home environments, performing tasks such as housework, educational assistance, and companionship. Professional service robots are used in commercial, medical, military, and industrial environments, performing specialized tasks such as surgical assistance, operating in hazardous environments, and public safety monitoring. They typically include sensors, actuators, control systems, and software that enable robots to operate independently with minimal or no human intervention.

[0003] Among them, the large-model optimized intelligent tour guide and photography robot involves a robot that integrates advanced AI optimization technology to improve the robot's performance and efficiency in tour guidance and photography tasks. By using large-model optimization, the robot can more accurately understand and respond to its environment, such as providing tour guidance at tourist attractions or taking high-quality photos of specific events. The main uses of this type of robot include providing information and entertainment, helping people better understand their surroundings, and also capturing and recording important visual information.

[0004] While existing service robots have been widely used in various fields, their adaptability and ability to provide personalized services in complex and dynamic environments remain limited. In particular, at tourist attractions with dense crowds, conventional service robots often cannot effectively respond to rapidly changing crowd dynamics and environmental variables, resulting in slow or even erroneous service responses, impacting the visitor experience. The lack of highly flexible emotional perception also makes it difficult for robots to precisely adjust their service methods to suit the moods and needs of different visitors, limiting their performance in personalized interactions. Furthermore, delayed information updates and poorly chosen photo opportunities can prevent robots from providing effective guide content and visual records at critical moments, reducing their practicality and appeal. Summary of the Invention

[0005] The purpose of the present invention is to solve the shortcomings of the prior art and propose an intelligent tour guide and photo taking robot based on large model optimization.

[0006] To achieve the above objectives, the present invention adopts the following technical solution: an intelligent tour guide and photography robot based on large model optimization includes a main body, characterized in that the main body is equipped with a robot control system, and the system includes:

[0007] The route optimization and dynamic adjustment module receives environmental variables and crowd dynamic data, performs real-time monitoring and analysis, calculates the route efficiency by comparing the congestion index of different routes, and selects the guided route based on the calculated path efficiency using a delay-minimizing strategy. It then dynamically adjusts the route based on real-time traffic changes to obtain the final adjusted guided route.

[0008] The emotion perception and interactive content adjustment module collects tourists' behavioral response data through sensors, calculates the emotion index by analyzing the data and combining it with the distribution of tourists on the final guided tour route after adjustment, adjusts the speaking speed and depth of the guide robot according to the calculated emotion index, determines the interactive content that satisfies the emotions and interests of different tourists, and obtains the adjusted guided tour content;

[0009] The real-time update module of the tour content performs an in-depth analysis of the information currently provided by the robot based on the adjusted tour content, performs information filtering and priority sorting, and adjusts the explanation content and display order of the tour robot in real time to obtain real-time updated tour information;

[0010] The photo timing and location selection module analyzes the real-time updated guide information and current lighting conditions, combines it with the landscape location data, performs time-location matching, calculates the photo time and location, and optimizes the robot's guide stop points and adjusts the photo angle based on the calculated photo time and location to obtain the optimal photo solution.

[0011] As a further solution of the present invention, the steps for calculating the path efficiency are:

[0012] Collect environmental and crowd dynamic data, including temperature, humidity, crowd flow and movement speed, and perform standardization.

[0013]

[0014] Calculate the deviation X of each data point from the mean std ,Generate standardized environment and crowd datasets, where,X,is the raw data collected by the sensor, μ,is the mean, and σ,is the standard deviation;

[0015] Based on the standardized environment and population data set, time series analysis was performed using the formula,

[0016]

[0017] Calculate the weighted average MA t , generate weighted sliding average monitoring data, where w i is the weight factor, W is the sum of the weight factors, X std,i represents the data points in the standardized environment and population dataset, and t represents the total number of data points;

[0018] According to the weighted moving average monitoring data, the congestion levels of different routes are compared and the formula is used.

[0019]

[0020] Calculate the path efficiency score E p , generate path efficiency integration data, where C E represents the congestion parameter of the route, β0 and β1 are model parameters, and exp(·) represents the natural exponential function.

[0021] As a further solution of the present invention, the steps for obtaining the optimized navigation path are:

[0022] Analyze the path efficiency integration data, combine the physical length of the path and the expected traffic conditions, and use the formula,

[0023]

[0024] Calculating Delay Score D r , generate expected delay data, where L is the route length, E p is the path efficiency score, γ is the adjustment coefficient;

[0025] According to the expected delay data, the optimal navigation path is selected using the formula,

[0026] R min =min(D r )

[0027] Get the optimal selection result of the navigation path, where R min Indicates the navigation path with the lowest delay score;

[0028] According to the optimal selection result of the navigation path, the path is dynamically adjusted in combination with real-time traffic changes, using the formula:

[0029] P opt =R min ×(1-α p ·T p )

[0030] Calculate the guide path score P opt , get the final adjusted navigation path, where α p is the adjustment coefficient, T p is the real-time traffic variation factor.

[0031] As a further solution of the present invention, the steps for calculating the emotion index are:

[0032] Based on multimodal sensors, we synchronously capture the facial expressions and voice tones of tourists and use the formula:

[0033]

[0034] Calculate the tourist data influence score D raw , generate a synchronized dataset of facial expressions and vocal tones, where x i and s i Represents the facial expression and tone data of the i-th tourist, n represents the number of tourists, a D and b D is the data weighting coefficient, which is used to balance the influence of facial expression and voice tone data;

[0035] Analyze the synchronous data set of facial expressions and voice tones, combine the distribution of tourists on the final guided path after adjustment, and associate them with the tourist locations. Use the formula:

[0036]

[0037] Calculate the position-weighted sentiment index E loc , generating position-weighted sentiment data, where l i is the location importance index of the i-th tourist, c D and d D is a regulatory factor based on position importance, D raw,i represents the data influence score of the i-th tourist;

[0038] Based on the position-weighted sentiment data, the formula is used,

[0039]

[0040] Calculate the overall emotional state score E of the tourist group total , generate the analysis results of tourists’ overall emotional state, where E loc,i represents the location-weighted sentiment index of the i-th tourist.

[0041] As a further solution of the present invention, the steps for obtaining the adjusted navigation content are:

[0042] Based on the analysis results of the tourists' overall emotional state, the speech speed parameters of the tour guide robot are adjusted using the formula:

[0043] R speed =γ R ln(E total +1)+δ R

[0044] Calculate the quantitative evaluation value R of speech speed speed , generate the adjusted speech rate parameters, where γ R and δ Ris the sensitivity parameter of speech rate adjustment, ln(E total +1) is used to smooth the impact of emotional indicators and avoid extreme values ​​affecting speech speed;

[0045] Based on the analysis results of the tourists' overall emotional state, the interaction depth of the tour guide robot is adjusted using the formula:

[0046]

[0047] Calculate the quantitative evaluation value R of the interaction depth depth , generates the adjusted interaction depth parameter, where λ R , κ, and μ R is the adjustment coefficient of interaction depth, To prevent overreaction;

[0048] The adjusted speech rate parameter and the adjusted interaction depth parameter are combined to customize the guide content, match different emotional levels, and deploy it to the guide robot to obtain the adjusted guide content.

[0049] As a further solution of the present invention, the steps for obtaining the real-time updated navigation information are:

[0050] Based on the adjusted navigation content, all the information currently provided by the navigation robot is collected and analyzed in depth, using the formula:

[0051]

[0052] Generate filtered content dataset D extracted , where c i represents the i-th content unit extracted from the robot content library, r i and j i is the weight of importance and urgency of content unit i, t i is the timeliness index of content unit i, and N represents the number of content units;

[0053] Based on the filtered content data set, filtering and prioritization are performed using the formula,

[0054]

[0055] Calculate the ranking level D of the navigation content sorted , generate the sorted navigation content dataset, where v i is the dynamic evaluation coefficient generated for each content item based on user feedback;

[0056] The sorted navigation content data set is used to dynamically display the navigation content on the user interface of the navigation robot in a descending order of priority, thereby obtaining real-time updated navigation information.

[0057] As a further solution of the present invention, the steps for calculating the photographing time point and location are as follows:

[0058] Based on the real-time updated guide information, the lighting data and landscape location are collected synchronously, and the formula is used.

[0059] Z={I Z , L Z , P Z}

[0060] Generate a comprehensive dataset, where I Z 、L Z and P Z Represent navigation information, light intensity and location coordinates respectively;

[0061] Based on the comprehensive data set, time-position matching is performed using the formula,

[0062]

[0063] Calculate the time-position matching score M, where l Mi represents the light intensity at position i, t Mi Recommended photo taking time for each location, T h is the total number of time units in a day, and m represents the number of locations;

[0064] Based on the time-position matching score, the optimal photographing time and position are selected, and the optimal photographing time point and position data are generated.

[0065] As a further solution of the present invention, the steps for obtaining the optimal photographing solution are:

[0066] Based on the optimal photo taking time and location data, the stopping point of the tour guide robot is optimized using the formula:

[0067] S=M max ×δ m

[0068] Calculate the optimized stay point score S and obtain the optimized robot stay point data, where δ m is the adjustment coefficient for terrain factors;

[0069] According to the optimized robot stop point data, adjust the shooting angle and use the formula:

[0070]

[0071] Calculate the adjusted shooting angle A, where θ and φ are the coefficients of the tilt angle and rotation angle respectively;

[0072] Combining the optimized robot stop point data and the adjusted photographing angle, a photographing plan is determined, a photographing operation is performed, and an optimal photographing plan is output.

[0073] Compared with the prior art, the advantages and positive effects of the present invention are:

[0074] In the present invention, dynamic data analysis is used to ensure the efficiency and practicality of route selection. The ability to respond to environmental changes in real time significantly improves the fluency and accuracy of the guided tour process, avoiding unnecessary delays. By analyzing the interactive content of tourists' behavioral response data, the robot can more accurately adjust the service content to match the tourists' emotions and interests, enhance the personalization of the service and the tourists' sense of participation. The real-time update of the guided tour content and the precise selection of the timing for taking photos further improve the timeliness and relevance of the information service, ensuring that tourists obtain valuable visual records and guided tour information at the best time. BRIEF DESCRIPTION OF THE DRAWINGS

[0075] Figure 1 A diagram showing the structure of the robot of the present invention;

[0076] Figure 2 is a robot flow chart of the present invention;

[0077] Figure 3 Flowchart for calculating path efficiency of the present invention;

[0078] Figure 4 This is a flowchart for obtaining the optimized navigation path of the present invention;

[0079] Figure 5 is a flow chart for calculating the emotion index of the present invention;

[0080] Figure 6 This is a flowchart for obtaining the adjusted navigation content of the present invention;

[0081] Figure 7 A flowchart of obtaining navigation information updated in real time according to the present invention;

[0082] Figure 8 This is a flow chart for calculating the time point and location of taking a photo in the present invention;

[0083] Figure 9 This is a flow chart for obtaining the optimal photographing solution of the present invention.

[0084] Legend:

[0085] 1. Subject. DETAILED DESCRIPTION

[0086] In order to make the purpose, technical solutions and advantages of the present invention more clearly understood, the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0087] In the description of the present invention, it should be understood that the terms "length," "width," "up," "down," "front," "back," "left," "right," "vertical," "horizontal," "top," "bottom," "inside," "outside," and the like, indicating positions or relationships, are based on the positions or relationships shown in the accompanying drawings and are intended only to facilitate the description of the present invention and simplify the description. They do not indicate or imply that the devices or elements referred to must have a specific orientation, be constructed, or operate in a specific orientation. Therefore, they should not be construed as limiting the present invention. Furthermore, in the description of the present invention, "plurality" means two or more, unless otherwise expressly and specifically defined.

[0088] See also Figure 1-2 The intelligent tour guide and photo taking robot based on large model optimization includes a main body 1, which is equipped with a robot control system. The system includes:

[0089] The route optimization and dynamic adjustment module receives environmental variables and crowd dynamic data, performs real-time monitoring and analysis, calculates the route efficiency by comparing the congestion index of different routes, and selects the guided route based on the calculated path efficiency using a delay-minimizing strategy. It then dynamically adjusts the route based on real-time traffic changes to obtain the final adjusted guided route.

[0090] The emotion perception and interactive content adjustment module collects tourists' behavioral response data through sensors, analyzes the data and combines it with the tourist distribution along the final adjusted guided tour path to calculate the emotion index. Based on the calculated emotion index, the guide robot's speech speed and depth are adjusted to determine the interactive content that meets the different emotions and interests of tourists, and obtain the adjusted guided tour content.

[0091] The real-time update module of the tour content conducts in-depth analysis of the information currently provided by the robot based on the adjusted tour content, performs information filtering and priority sorting, and adjusts the explanation content and display order of the tour robot in real time to obtain real-time updated tour information;

[0092] The photo timing and location selection module analyzes the real-time updated guide information and current lighting conditions, combines it with landscape location data, performs time-position matching, calculates the photo time and location, and optimizes the robot's guide stop points and adjusts the photo angle based on the calculated photo time and location to obtain the optimal photo plan.

[0093] The adjusted final tour path includes the route selection results, route adjustment records and route confirmation results. The adjusted tour content includes the speech speed setting results, content depth adjustment records and content matching results. The real-time updated tour information includes the explanation content display order sorting results and information update records. The optimal photo plan includes the photo time, photo location and photo stop point selection results.

[0094] See also Figure 3 , the calculation steps of path efficiency are:

[0095] Collect environmental and crowd dynamic data, including temperature, humidity, crowd flow and movement speed, and perform standardization.

[0096]

[0097] Calculate the deviation X of each data point from the mean std ,Generate standardized environment and crowd datasets, where,X,is the raw data collected by the sensor, μ,is the mean, and σ,is the standard deviation;

[0098] Based on the standardized environment and population data set, time series analysis is performed using the formula,

[0099]

[0100] Calculate the weighted average MA t , generate weighted sliding average monitoring data, where w i is the weight factor, W is the sum of the weight factors, X std,i represents the data points in the standardized environment and population dataset, and t represents the total number of data points;

[0101] According to the monitoring data of weighted sliding average, the congestion degree of different routes is compared and the formula is used.

[0102]

[0103] Calculate the path efficiency score E p , generate path efficiency integration data, where C E represents the congestion parameter of the route, β0 and β1 are model parameters, and exp(·) represents the natural exponential function.

[0104] Taking temperature data as an example, the collected temperature data (X) is as follows (unit: degrees Celsius): [22, 23, 21, 25, 24].

[0105] Calculate the mean μ:

[0106]

[0107] Calculate the standard deviation σ:

[0108]

[0109] Normalize data:

[0110]

[0111]

[0112] The standardized data set [-0.71, 0, -1.42, 1.42, 0.71] represents the deviation of each data point from the mean (in standard deviations), which helps to reduce the impact of data dimension in subsequent analysis.

[0113] Weight factor w i Set it to [0.1, 0.2, 0.4, 0.2, 0.1] and perform weighted average calculation on the above standardized data.

[0114] The total weight W:

[0115] W=0.1+0.2+0.4+0.2+0.1=1

[0116] Calculate the weighted average MA:

[0117]

[0118] MA=-0.071+0-0.568+0.284+0.071

[0119] MA=-0.284

[0120] The weighted moving average result of -0.284 indicates that within the current time window, the smaller negative numbers show a slight negative deviation from the average.

[0121] On a specific route, the congestion parameter C E is 0.5, and the model parameters β0=0.2 and β1=1.5 are set.

[0122] Substitute into the formula to calculate the efficiency score E p :

[0123]

[0124] A path efficiency score of 0.721 indicates that under the current congestion parameters, the traffic efficiency of this path is 72.1%. A higher score indicates that the path is relatively less congested.

[0125] See also Figure 4 , the steps to obtain the optimized navigation path are:

[0126] Analyze the path efficiency integration data, combine the physical length of the path and the expected traffic conditions, and use the formula,

[0127]

[0128] Calculating Delay Score D r , generate expected delay data, where L is the route length, E p is the path efficiency score, γ is the adjustment coefficient;

[0129] According to the expected delay data, the optimal navigation path is selected using the formula:

[0130] R min =min(D r )

[0131] Get the optimal selection result of the navigation path, where R min Indicates the navigation path with the lowest delay score;

[0132] According to the optimal selection result of the guided route, the route is dynamically adjusted in combination with real-time traffic changes. The formula is used.

[0133] P opt =R min ×(1-α p ·T p )

[0134] Calculate the guide path score P opt , get the final adjusted navigation path, where α p is the adjustment coefficient, T p is the real-time traffic variation factor.

[0135] L: Length of the route, 5 kilometers actually measured.

[0136] γ: Adjustment coefficient, set to 1.2 to emphasize the impact of efficiency score.

[0137] Combined with the above calculation results, calculate the delay score D r :

[0138]

[0139] A delay score of 7.41 indicates that under the current route efficiency, taking into account length and efficiency factors, the average delay per kilometer is expected to be approximately 1.482 minutes (total delay time divided by total length), indicating that the route faces a high risk of delay.

[0140] The delay scores of other paths are calculated as [11.0, 12.5, 6, 7.41], including the current path score.

[0141] Select minimum delay rating:

[0142] R min =min(11.0,12.5,6,7.41)

[0143] R min =6

[0144] The optimal guide route selection result 6 indicates that the route has the lowest delay score in the comparison and is selected as the preferred route for the guide. It is expected to have a lower time loss and is suitable for the guide.

[0145] α p : Adjustment coefficient, set to 0.05, indicating the sensitivity of real-time adjustment.

[0146] T p : Real-time traffic change factor, set to 0.1, reflecting the increase or decrease of real-time traffic delays.

[0147] Calculate the adjusted tour path score:

[0148] P opt =6×(1-0.05·0.1)=6×(1-0.005)

[0149] P opt =6×0.995=5.97

[0150] The adjusted final route score of 5.97 indicates that the route efficiency improved slightly after fine-tuning the actual route selection to account for real-time traffic changes. This number indicates that the route remained optimal despite real-time factors, but the subtle adjustments provided a more accurate prediction of the route duration.

[0151] See also Figure 5 , the calculation steps of the sentiment index are:

[0152] Based on multimodal sensors, we synchronously capture the facial expressions and voice tones of tourists and use the formula:

[0153]

[0154] Calculate the tourist data influence score D raw , generate a synchronized dataset of facial expressions and vocal tones, where x i and s i Represents the facial expression and tone data of the i-th tourist, n represents the number of tourists, a D and b D is the data weighting coefficient, which is used to balance the influence of facial expression and voice tone data;

[0155] Analyze the synchronous data set of facial expressions and voice tones, combine the distribution of tourists on the final adjusted guided path, and associate it with the tourist location.

[0156]

[0157] Calculate the position-weighted sentiment index E loc , generating position-weighted sentiment data, where l i is the location importance index of the i-th tourist, c D and d D is a regulatory factor based on position importance, D raw,i represents the data influence score of the i-th tourist;

[0158] Based on the position-weighted sentiment data, the formula is used:

[0159]

[0160] Calculate the overall emotional state score E of the tourist group total , generate the analysis results of tourists’ overall emotional state, where E loc,i represents the location-weighted sentiment index of the i-th tourist.

[0161] For a small sample of tourists (n=3), the following data were collected:

[0162] Facial expression value x i =[0.5, 0.6, 0.4].

[0163] Tone value s i =[0.7, 0.5, 0.6].

[0164] Weighting coefficient, set a D = 0.6 and b D =0.4.

[0165] Substituting into the formula, we get:

[0166] D raw,1 =0.6×0.5+0.4×0.7=0.3+0.28=0.58

[0167] D raw,2 =0.6×0.6+0.4×0.5=0.36+0.2=0.56

[0168] D raw,3 =0.6×0.4+0.4×0.6=0.24+0.24=0.48

[0169] These values ​​(0.58, 0.56, 0.48) represent the overall influence score for each tourist data, with higher scores indicating stronger emotional responses.

[0170] Location importance i =[1, 0.8, 0.9].

[0171] Adjustment factor, determine c D =0.7 and d D =0.3.

[0172] Substituting into the formula, we get:

[0173] E loc,1 =0.7×0.58+0.3×1=0.406+0.3=0.706

[0174] E loc,2 =0.7×0.56+0.3×0.8=0.392+0.24=0.632

[0175] E loc,3 =0.7×0.48+0.3×0.9=0.336+0.27=0.606

[0176] These values ​​(0.706, 0.632, 0.606) represent the sentiment indicators after considering tourists’ locations, which can help optimize the tour content to better adapt to tourists’ mood changes in different locations.

[0177] Combining the above data, we can get:

[0178]

[0179] E total =0.648, which represents the average emotional state of the entire tourist group and is used to adjust the tour content to ensure that the content meets the emotional needs of tourists.

[0180] See also Figure 6 , the steps to obtain the adjusted guide content are:

[0181] Based on the analysis results of tourists' overall emotional state, the speech speed parameters of the tour guide robot are adjusted using the formula:

[0182] R speed =γ R ln(E total +1)+δ R

[0183] Calculate the quantitative evaluation value R of speech speed speed , generate the adjusted speech rate parameters, where γ R and δ Ris the sensitivity parameter of speech rate adjustment, ln(E total +1) is used to smooth the impact of emotional indicators and avoid extreme values ​​affecting speech speed;

[0184] Based on the analysis results of tourists' overall emotional state, the interaction depth of the tour guide robot is adjusted using the formula:

[0185]

[0186] Calculate the quantitative evaluation value R of the interaction depth depth , generates the adjusted interaction depth parameter, where λ R , κ, and μ R is the adjustment coefficient of interaction depth, To prevent overreaction;

[0187] Combined with the adjusted speech speed parameters and the adjusted interaction depth parameters, the guided tour content is customized to match different emotional levels, and deployed to the guided tour robot to obtain the adjusted guided tour content.

[0188] Given the overall emotional state E total =0.648.

[0189] Setting γ R =2.5 and δ R =0.5, substitute into the formula and calculate:

[0190] R speed =2.5×ln(0.648+1)+0.5

[0191] R speed ≈2.5×0.499+0.5=1.2475+0.5=1.7475

[0192] R speed =1.7475 represents the adjusted speech speed quantitative evaluation value, which can be used to set the speaking speed of the tour guide robot. A lower positive value reflects a moderate increase in the speaking speed to accommodate moderately excited tourists.

[0193] Setting λ R =3, κ = 1.5 and μ R =0.2, then:

[0194] R depth =3×(1-e -1.5×0.648 )+0.2

[0195] R depth ≈3×(1-0.378)+0.2=3×0.622+0.2=2.066

[0196] R depth=2.066 represents the adjusted interaction depth parameter. The higher value reflects the enhanced interaction depth to provide richer and deeper information to accommodate tourists' higher emotional participation.

[0197] Finally, the guided tour content is customized by combining the adjusted speech speed parameters and the adjusted interaction depth parameters.

[0198] See also Figure 7 , the steps to obtain the real-time updated guide information are:

[0199] Based on the adjusted tour content, all the information currently provided by the tour guide robot is collected and analyzed in depth, using the formula:

[0200]

[0201] Generate filtered content dataset D extracted , where c i represents the i-th content unit extracted from the robot content library, r i and j i is the weight of importance and urgency of content unit i, t i is the timeliness index of content unit i, and N represents the number of content units;

[0202] Based on the filtered content data set, filtering and prioritization are performed using the formula,

[0203]

[0204] Calculate the ranking level D of the navigation content sorted , generate the sorted navigation content dataset, where v i is the dynamic evaluation coefficient generated for each content item based on user feedback;

[0205] Using the sorted guide content dataset, the guide content is dynamically displayed on the user interface of the guide robot in a descending order of priority, so as to obtain real-time updated guide information.

[0206] Three content units were collected, and the specific parameters of each content were as follows:

[0207] c1, c2, c3 = [information 1, information 2, information 3].

[0208] r1, r2, r3 = [0.8, 0.5, 0.9], based on the user feedback score.

[0209] j1, j2, j3 = [0.2, 0.8, 0.1], reflecting the urgency of the information.

[0210] t1, t2, t3 = [0.9, 0.4, 0.7], reflecting the relevance of information to the current event.

[0211] Substituting into the formula, we get:

[0212] D extracted,1 =0.8×c1+0.2×0.9=0.98c1

[0213] D extracted,2 =0.5×c2+0.8×0.4=0.82c2

[0214] D extracted,3 =0.9×c3+0.1×0.7=0.97c3

[0215] Lifelong D extracted It is a set of content units with adjusted weights to ensure that the priority and timeliness of information meet the current navigation needs.

[0216] The evaluation coefficient of each content item generated based on user feedback is as follows:

[0217] v1, v2, v3=[1.2, 0.5, 2.0]

[0218] Substituting into the formula, we get:

[0219]

[0220] D sorted Represents the content ranking after adjustment of the evaluation coefficient, ensuring that the display order prioritizes user preferences and content popularity. From the results, we can see that the display order is (c3, c1, c2).

[0221] See also Figure 8 , the calculation steps for the photo taking time and location are:

[0222] Based on the real-time updated guide information, the lighting data and landscape location are collected synchronously, and the formula is used.

[0223] Z={I Z , L Z , P Z}

[0224] Generate a comprehensive dataset, where I Z 、L Z and P Z Represent navigation information, light intensity and location coordinates respectively;

[0225] Based on the comprehensive data set, time-position matching is performed using the formula,

[0226]

[0227] Calculate the time-position matching score M, where l Mi represents the light intensity at position i, t Mi Recommended photo taking time for each location, T h is the total number of time units in a day, and m represents the number of locations;

[0228] Based on the time-location matching score, the optimal photo taking time and location are selected to generate the optimal photo taking time point and location data.

[0229] At a certain scenic spot, the returned guide information I Z It is the historical information about the point, the light intensity L measured by the light sensor Z is 5000 lux, position coordinate P Z is the GPS coordinate (35.6895, 139.6917), then:

[0230] Z = {historical information, 5000, (35.6895, 139.6917)}

[0231] Data were collected at three locations, with light intensities of 4500 Iux, 5000 Iux, and 4800 Iux, respectively. The maximum light intensity was max(l M )=5000lux.

[0232] Recommended photo taking time t Mi They are the 10th hour, the 12th hour, and the 11th hour respectively.

[0233] T h The total number of time units in a day is 24 hours.

[0234] Substitute into the formula to calculate the matching score:

[0235]

[0236] M≈0.81×0.4167+1×0.5+0.9216×0.4583

[0237] M=0.337527+0.5+0.42236928≈1.26

[0238] The resulting matching score, M, represents the optimal combination of lighting and time for each attraction. It reflects a comprehensive evaluation of the lighting conditions and recommended photo times at each attraction at a specific time. A higher score indicates a closer match between lighting conditions and photo timing at that particular time and location, and is used to select the optimal time and location for photo shooting.

[0239] See also Figure 9 , the steps to obtain the optimal photo-taking solution are:

[0240] Based on the optimal photo taking time and location data, the guide robot's stopping point is optimized using the formula:

[0241] S=M max ×δ m

[0242] Calculate the optimized stay point score S and obtain the optimized robot stay point data, where δ m is the adjustment coefficient for terrain factors;

[0243] According to the optimized robot stop point data, adjust the shooting angle and use the formula:

[0244]

[0245] Calculate the adjusted shooting angle A, where θ and φ are the coefficients of the tilt angle and rotation angle respectively;

[0246] Combined with the optimized robot stop point data and the adjusted shooting angle, the shooting plan is determined, the shooting operation is executed, and the optimal shooting plan is output.

[0247] According to the calculation, M max =1.26, adjustment coefficient δ considering terrain factors m =1.1, indicating that the terrain in this area is relatively open and suitable for staying and taking photos.

[0248] Substituting into the formula, we get:

[0249] S=1.26×1.1=1.386

[0250] 1.386 represents the optimized stay point score, which directly reflects the suitability of the best stay point after considering the terrain adjustment.

[0251] Combined with the above calculation result S = 1.386, we get the tilt angle θ = 30° and the rotation angle coefficient φ = 0.9, then:

[0252]

[0253] A = 1.44° represents the adjusted shooting angle. This angle is further adjusted based on the optimized stop point, taking into account the optimal viewing angle and lighting conditions of the landscape. This adjusted angle optimizes the landscape display during photography, enhancing the visual impact of the photo.

[0254] The above are merely preferred embodiments of the present invention and do not limit the present invention in any other form. Any technician familiar with the profession may use the technical content disclosed above to change or modify it into an equivalent embodiment with equivalent changes and apply it to other fields. However, any simple modification, equivalent change and modification made to the above embodiment based on the technical essence of the present invention without departing from the content of the technical solution of the present invention shall still fall within the scope of protection of the technical solution of the present invention.

Claims

1. An intelligent tour guide and photo taking robot based on large model optimization, comprising a main body (1), characterized in that: The main body (1) is equipped with a robot control system, which includes: The route optimization and dynamic adjustment module receives environmental variables and crowd dynamic data, conducts real-time monitoring and analysis, calculates the route efficiency by comparing the congestion index of different routes, selects the guided route based on the calculated results of the route efficiency, dynamically adjusts the route based on real-time traffic changes, and obtains the final guided route after adjustment; The emotion perception and interactive content adjustment module collects tourists' behavioral response data through sensors, calculates the emotion index by analyzing the data and combining it with the distribution of tourists on the final guided tour path after adjustment, adjusts the speech speed and depth of the guided tour robot according to the calculation result of the emotion index, determines that the interactive content satisfies the emotions and interests of different tourists, and obtains the adjusted guided tour content; The real-time updating module of the guide content performs in-depth analysis on the information currently provided by the robot based on the adjusted guide content, performs information filtering and priority sorting, adjusts the explanation content and display order of the guide robot in real time, and obtains real-time updated guide information; The photo timing and location selection module analyzes the real-time updated guide information and current lighting conditions, combines the landscape location data, performs time-location matching, calculates the photo time and location, optimizes the robot guide stop point based on the photo time and location calculation results, adjusts the photo angle, and obtains the optimal photo solution.

2. The intelligent tour guide and photo taking robot based on large model optimization according to claim 1, characterized in that: The calculation steps of the path efficiency are: Collect environmental and crowd dynamic data, including temperature, humidity, crowd flow and movement speed, and standardize them using formulas. Calculate the deviation of each data point from the mean X std , Generate standardized environment and crowd datasets, where X is the raw data collected by the sensor, μ is the mean, and σ is the standard deviation; Based on the standardized environment and population data set, time series analysis was performed using the formula, Calculate the weighted average MA t , generate weighted moving average monitoring data, where w i is the weight factor, W is the sum of the weight factors, X std,i represents the data points in the standardized environment and population dataset, and t represents the total number of data points; According to the monitoring data of the weighted moving average, the congestion degree of different routes is compared, and the formula is used: Calculate the path efficiency score E p , generate path efficiency integration data, where C E represents the congestion parameter of the route, β0 and β1 are model parameters, and exp(·) represents the natural exponential function.

3. The intelligent tour guide and photo taking robot based on large model optimization according to claim 2, characterized in that: The steps for obtaining the optimized navigation path are: Analyze the path efficiency integration data, combine the physical length of the path and the expected traffic conditions, and use the formula, Calculate Delay Score D r , generate expected delay data, where L is the route length, E p is the path efficiency score, γ is the adjustment coefficient; According to the expected delay data, the optimal navigation path is selected using the formula, R min =min(D r ) Get the optimal selection result of the navigation path, where R min Indicates the navigation path with the lowest delay score; According to the optimal selection result of the navigation path, the path is dynamically adjusted in combination with real-time traffic changes, using the formula: P opt =R min ×(1-a p ·T p ) Calculate the guide path score P opt , get the adjusted final navigation path, where α p is the adjustment factor, T p is the real-time traffic variation factor.

4. The intelligent tour guide and photo taking robot based on large model optimization according to claim 3 is characterized in that: The calculation steps of the sentiment index are: Based on multimodal sensors, the facial expressions and voice tones of tourists are captured synchronously, and the formula is adopted. Calculate the tourist data influence score D raw , generate a synchronized dataset of facial expressions and vocal intonations, where x i and i represents the facial expression and tone of voice data of the i-th tourist, n represents the number of tourists, a D and b D is the data weighting coefficient, which is used to balance the influence of facial expression and voice tone data; Analyze the synchronous data set of facial expressions and intonations, combine the distribution of tourists on the adjusted final guided path, and associate with the tourist positions, using the formula, Calculate the position-weighted sentiment index E loc , generating position-weighted sentiment data, where l i is the location importance index of the i-th tourist, c D and d D is a moderating factor based on position importance, D raw,i represents the data influence score of the i-th tourist; Based on the position-weighted sentiment data, the formula is used, Calculate the overall emotional state score E of the tourist group total , generate the analysis results of tourists’ overall emotional state, where E loc,i represents the location-weighted sentiment index of the i-th tourist.

5. The intelligent tour guide and photo taking robot based on large model optimization according to claim 4 is characterized in that: The steps for obtaining the adjusted navigation content are as follows: Based on the analysis results of the tourists' overall emotional state, the speech speed parameters of the tour guide robot are adjusted using the formula: R speed =c R ·ln(E total +1)+d R Calculate the quantitative evaluation value R of speech speed speed , generate the adjusted speech rate parameters, where γ R and δ R is the sensitivity parameter of speech rate adjustment, ln(E total +1) is used to smooth the impact of emotional indicators and avoid extreme values ​​affecting speech speed; Based on the analysis results of the tourists' overall emotional state, the interactive depth of the tour guide robot is adjusted using the formula: Calculate the interaction depth quantitative evaluation value R depth , generating the adjusted interaction depth parameter, where λ R , κ and μ R is the adjustment factor of the interaction depth, To prevent overreaction; Combined with the adjusted speech speed parameter and the adjusted interaction depth parameter, the guided tour content is customized to match the different emotional levels, and deployed to the guided tour robot to obtain the adjusted guided tour content.

6. The intelligent tour guide and photo taking robot based on large model optimization according to claim 5, characterized in that: The steps for obtaining the real-time updated navigation information are as follows: Based on the adjusted navigation content, all the information currently provided by the navigation robot is collected and analyzed in depth, using the formula: Generate the filtered content dataset D extracted , where c i represents the i-th content unit extracted from the robot content library, r i and j i is the weight of importance and urgency of content unit i, t i is the timeliness index of content unit i, and N represents the number of content units; Based on the filtered content data set, filtering and prioritization are performed, using the formula, Calculate the ranking level D of the navigation content sorted , generate a sorted navigation content dataset, where v i is the dynamic evaluation coefficient generated for each content item based on user feedback; The sorted navigation content data set is used to dynamically display the navigation content on the user interface of the navigation robot in a descending order of priority, so as to obtain the navigation information updated in real time.

7. The intelligent tour guide and photo taking robot based on large model optimization according to claim 6, characterized in that: The calculation steps of the photographing time point and position are as follows: Based on the real-time updated guide information, the lighting data and landscape location are collected synchronously, and the formula is used: Z={I Z ,L Z ,P Z } Generate a comprehensive dataset, where I Z , L Z and P Z Represent navigation information, light intensity and location coordinates respectively; Based on the comprehensive data set, time-position matching is performed using the formula, Calculate the time-position matching score M, where l Mi represents the light intensity at position i, t Mi Recommended photo taking time for each location, T h is the total number of time units in a day, and m represents the number of locations; Based on the time-position matching score, the optimal photographing time and position are selected to generate the optimal photographing time point and position data.

8. The intelligent tour guide and photo taking robot based on large model optimization according to claim 7, characterized in that: The steps for obtaining the optimal photographing solution are as follows: Based on the optimal photo-taking time and location data, the stop point of the tour guide robot is optimized using the formula: S=M max ×δ m Calculate the optimized stay point score S and obtain the optimized robot stay point data, where δ m is the adjustment coefficient for terrain factors; According to the optimized robot stop point data, the camera angle is adjusted and the formula is used. Calculate the adjusted shooting angle A, where θ and φ are the coefficients of the tilt angle and the rotation angle respectively; Combined with the optimized robot stop point data and the adjusted shooting angle, a shooting plan is determined, a shooting operation is performed, and an optimal shooting plan is output.

Images