A transparent capability trust presentation method for unmanned aerial vehicles

Abstract

The application belongs to the technical field of computers, and particularly relates to a transparent capability trust presentation method for unmanned aerial vehicles. A three-dimensional trust interaction framework based on a trust calibration model is constructed, with capability trust, integrity trust and goodwill trust serving as design pillars. A multi-source fusion calculation model of flight readiness index is designed, which integrates three types of data, i.e. meteorological conditions, airspace density and vehicle health, adopts a weighted linear combination model to calculate the flight readiness index and presents it in the form of a percentage system. An active transparent disclosure mechanism of delay probability is established, the delay probability is estimated based on a logistic regression model, and when the delay probability exceeds a preset threshold, an active notification process is triggered to broadcast the delay reason in the first person and push an alternative routing option. Gradual information disclosure control based on flight phases is realized, the flight phase is taken as an information quantity constraint condition, the passenger's working memory capacity is taken as an upper limit benchmark, full information is allowed to be presented in the cruising phase, only necessary safety information is retained in the takeoff and landing phases, and the minimum emergency mode is forced to enter in the case of abnormal alarm. A situation-adaptive cabin environment control algorithm is developed, the target color temperature is calculated according to the local time of flight, and dynamic balance between circadian rhythm lighting and personalized preference is realized. The application can effectively avoid the mis-calibration state of insufficient trust or excessive trust by establishing a complete three-dimensional trust calibration framework, realize the transparent presentation of flight state information, establish an active transparent disclosure mechanism of abnormal state, solve the technical problem that the existing scheme ignores the passenger's cognitive load limit, improve the personalized experience of the cabin environment, and strengthen the passenger's trust perception of the autonomous flight system.

Description

Technical Field

[0001] This application belongs to the field of computer technology, specifically relating to a transparent capability trust presentation method for unmanned aerial vehicles. Background Technology

[0002] With the accelerating pace of global urbanization and the rapid development of intelligent transportation technologies, urban air mobility, as a novel three-dimensional travel paradigm, is attracting widespread attention from academia and industry. Electric vertical takeoff and landing (EVTOL) aircraft, as the core transportation tool of urban air mobility systems, are considered a potential solution to the pain points of short-distance commuting in megacities due to their advantages such as strong vertical takeoff and landing capabilities, low noise levels, and relatively controllable operating costs. The commercialization of EVTOL aircraft is accelerating, and they are expected to play a role in connecting core urban areas with surrounding city nodes in future urban agglomeration transportation networks. However, the successful implementation of urban air mobility systems depends not only on the engineering reliability of the aircraft themselves, but also on the public's psychological acceptance and level of trust in autonomous flight systems. For ordinary passengers who have never encountered unmanned aerial vehicles before, boarding an aircraft without a human pilot for daily commuting presents an unprecedented cognitive and emotional challenge. Establishing and maintaining appropriate trust in autonomous flight systems has become a core bottleneck restricting the large-scale commercial application of urban air mobility.

[0003] In urban air mobility scenarios, the intelligent cockpit, as the sole direct interface between passengers and autonomous flight systems, undertakes multiple functions including information transmission, status presentation, and environmental control, making it a key touchpoint influencing passengers' trust perception and flight experience. Existing research on urban air mobility primarily focuses on technical aspects such as aircraft aerodynamic design, avionics system optimization, and airspace management mechanisms, while systematic research on the trust establishment and calibration mechanisms between passengers and autonomous flight systems is relatively scarce. The design of the intelligent cockpit's human-machine interface not only needs to meet basic functional information presentation requirements but also needs to strike a balance between functional efficiency and emotional experience, systematically responding to passengers' cognitive anxiety, emotional fluctuations, and social needs when facing autonomous flight systems. Passenger trust in autonomous flight systems is not a single-dimensional psychological variable but involves complex constructs across multiple levels, including judgments of the system's ability to complete tasks, assessments of the system's behavioral consistency and information transparency, and perceptions of whether the system acts in the passenger's best interests.

[0004] Existing technologies have significant shortcomings in the human-machine interaction design of intelligent cockpits for urban air mobility scenarios. First, most existing solutions simplify the trust issue to a simple information display problem, neglecting the multidimensional constructive nature of trust and lacking a theoretical framework for systematically designing the interface from different dimensions such as capability trust, integrity trust, and goodwill trust. Second, existing cockpit interaction systems generally lack a transparent presentation mechanism for flight status information. Passengers struggle to obtain key capability information such as system operating status, performance margins, and sensor health during flight, making it difficult for them to accurately assess the actual capabilities of the autonomous flight system and easily leading to misalignment states of insufficient or excessive trust. Third, existing solutions have flaws in the design of information disclosure mechanisms for abnormal flight conditions. When encountering situations such as route changes, predicted delays, or weather changes, the system often fails to establish a proactive, timely, and transparent information communication process, making it difficult to effectively alleviate passenger anxiety caused by information asymmetry. Finally, the design of existing cockpit environmental control systems and interfaces fails to fully consider the special psychological needs of first-time flight passengers, lacking specific interaction strategies designed for the critical trust-building window during the initial flight phase. This makes it difficult to eliminate the negative impact of the first flight experience on passengers' willingness to continue using the system. The aforementioned technical deficiencies collectively constitute the core obstacle restricting the improvement of trust experience in intelligent cockpits for urban air transportation, and a systematic solution is urgently needed from the perspective of theoretical framework construction and interaction strategy innovation.

[0005] The purpose of this invention is to provide a transparent capability trust presentation method for unmanned aerial vehicles, which can effectively solve the problems in the background art.

[0006] To achieve the above objectives, the technical solution adopted by the present invention is as follows: A transparent capability trust presentation method for unmanned aerial vehicles (UAVs) includes the following specific steps: Step 1: Constructing a three-dimensional trust interaction framework based on a trust calibration model: Capability trust, integrity trust, and goodwill trust are used as the three pillars of the design. A trust calibration mechanism is established between passengers and the autonomous flight system. The capability trust dimension is responded to through a real-time flight status visualization module, the integrity trust dimension is responded to through an abnormal status proactive notification interaction process, and the goodwill trust dimension is responded to through a personalized context perception module. Step 2: Designing a multi-source fusion calculation model for the flight readiness index: Integrating three types of data—meteorological conditions, airspace density, and aircraft health—a weighted linear combination model is used to calculate the flight readiness index. A min-max linear scaling method is used to normalize each sub-score, and the final score is presented to the passenger application interface in percentage form. Step 3: Establishing a main... Dynamic transparent disclosure mechanism: Based on a logistic regression model, the delay probability is estimated. The flight readiness index, remaining time until the scheduled departure time, airspace traffic density, and historical flight delay rates are used as input features. When the delay probability exceeds a preset threshold, an active notification process is triggered, broadcasting the cause of the delay in the first person and pushing alternative route options. Step 4 implements progressive information disclosure control based on flight phase: The flight phase is used as an information volume constraint, and the passenger's working memory capacity is used as the upper limit benchmark. Full information is allowed to be presented during the cruise phase, while only flight status and safety prompts are retained during takeoff and landing phases. When an abnormal alarm is triggered, the system is forced to enter a minimized emergency mode. Step 5 develops a scenario-adaptive cockpit environment control algorithm: The target color temperature is calculated based on the local flight time. When there is a passenger's personal color temperature preference, a weighted fusion output is used to achieve a dynamic balance between daytime rhythm lighting and personalized preferences.

[0007] Preferably, the design of the capability trust dimension in step 1 is centered on the real-time flight status visualization module. The system continuously and clearly presents system performance data, including sensor health, redundancy status and flight envelope margin. The information presentation uses common language instead of technical terms. The default mode uses a green indicator to intuitively present whether the system is operating normally, and provides real-time traffic summary in easy-to-understand natural language.

[0008] Preferably, the design of the integrity and trust dimension in step 1 is centered on real-time broadcasting of route changes and proactive notification of abnormal status. When the system encounters route changes, delay predictions, or weather changes, it establishes a proactive, timely, and transparent information communication process. The cause of the abnormality is explained in the first person to ensure that the anxiety of passengers caused by information asymmetry is effectively alleviated.

[0009] Preferably, the design of the benevolent trust dimension in step 1 is based on the cross-flight preference memory file and the soft takeoff emotional soothing sequence. The system remembers passenger preferences and actively adjusts the cabin environment. The soothing sound effect in both ear canals is activated in the first minute of flight. The personalized video briefing in the application can be viewed before departure. The augmented reality exit label in the cabin and the pre-flight safety briefing complete the safety education in a non-alarming tone.

[0010] Preferably, in step 2, the weighted linear combination model of the flight readiness index adopts a preset weight configuration, and the sum of the three weights is a preset total value. The system sets multiple thresholds. When the displayed value of the flight readiness index is greater than or equal to the first preset threshold, it is marked as ready and identified in green. When the displayed value is in the range between the first preset threshold and the second preset threshold, it is marked as feasible and prompts the user to pay attention to the weather. When the displayed value is lower than the second preset threshold, the pre-flight manual review process is triggered and the delay probability estimate is actively pushed to the passenger.

[0011] Preferably, in step 2, the meteorological score is calculated by combining three factors: wind speed, visibility, and precipitation probability, and a scoring method normalized to a preset range is used; the airspace density score is based on real-time traffic data and is processed by reverse normalization, with a higher score for lower traffic density; the aircraft health score is output by the airborne health management system, which integrates the propulsion system status, battery management system status, and structural integrity monitoring data.

[0012] Preferably, the input feature vector of the logistic regression model for the delay probability in step 3 includes the flight readiness index, the remaining time until the scheduled departure time, the current airspace traffic density, and the historical delay rate of the same route during the same period. The model coefficients are obtained by offline training from historical operating data and are iteratively updated with new data at a preset period. When the delay probability is greater than or equal to the preset probability threshold, the intelligent voice assistant announces the cause of the delay and the estimated waiting time in the first person and pushes alternative route options, conveying the uncertainty to passengers in a transparent manner.

[0013] Preferably, the calculation of passenger working memory capacity in step 4 is based on information chunking theory. The number of chunks is a preset number of chunks, the number of alternative states for each chunk is a preset number of states, and the calculated upper limit of capacity is a preset upper limit of capacity. The load factor is dynamically assigned according to the flight phase. The first preset value is used during the cruise phase, the second preset value is used during the takeoff and landing phases, and the load factor is forcibly reduced to the third preset value when an abnormal alarm is triggered.

[0014] Preferably, the information architecture design for the progressive information disclosure in step 4 adopts a hierarchical presentation strategy. The default mode only presents a simplified state to reduce the cognitive load on passengers. In-depth information is available to users as needed through secondary pages or extended entry points, reserving a path for users with proactive information needs to access in-depth information. This strategy ensures a balance between the accessibility of safety information and the control of passenger cognitive load.

[0015] Preferably, in step 5, the color temperature adjustment of the full-spectrum lighting system uses a preset neutral white reference color temperature as the neutral white reference color temperature, and the adjustment range is a preset color temperature adjustment range. During the day, the high color temperature reaches the preset high color temperature threshold, and at night, the low color temperature drops to the preset low color temperature threshold. The system automatically calculates the target color temperature according to the local flight time and the sinusoidal function law. When there is a passenger's personal color temperature preference, a preset fusion weight is used for weighted fusion output, with the passenger's historical preference as the main factor and the daytime rhythm algorithm as the auxiliary factor.

[0016] Preferably, it also includes a touch interaction optimization mechanism under vibration environment: the cabin acceleration sensor samples the root mean square value of the three-axis acceleration vector in real time. When the root mean square value exceeds the preset acceleration threshold, the minimum touch target width of the safety-critical control is automatically expanded from the first preset pixel width to the second preset pixel width, compressing the touch difficulty index to below the preset difficulty index upper limit, and ensuring that the touch response time does not exceed the preset response time upper limit.

[0017] Preferably, this also includes the entire service experience design: before flight, safety education is completed through personalized video briefings in the application and augmented reality exit markings in the cabin; during the first minute of flight, a flexible takeoff sequence is used to respond to the peak of physiological arousal; upon arrival, a connection reservation card is pushed out and a post-trip satisfaction score is presented in real time when emotions are at their peak, thus completing the feedback loop.

[0018] Preferably, it also includes a privacy-layered architecture design: personal biometric data is stored in an encrypted passenger self-control file, and aircraft performance and flight path data are retained independently in a de-identified form, which meets regulatory compliance requirements while ensuring the privacy protection of passengers' personal data.

[0019] Preferably, it also includes a collaboration mechanism between ground service personnel and the system: ground service personnel receive real-time updates from the background dispatch through a dedicated operation application, the maximum response delay for operation alarms is set to the preset maximum response delay, and mandatory read receipts are set to ensure the timeliness of information transmission; when the intelligent system detects passenger panic, a manual takeover protocol is initiated to ensure the controllability and continuity of passenger experience in service degradation scenarios.

[0020] Compared with the prior art, the present invention has the following beneficial effects: 1. Establish a systematic trust calibration design framework Breaking through the limitations of single-dimensional trust design: This invention uses three dimensions—capability trust, integrity trust, and goodwill trust—as design pillars to construct a complete trust calibration interaction framework. It solves the technical defect of existing solutions that simplify the trust issue to a single information display problem, enabling passengers' trust perception of the autonomous flight system to cover multiple levels such as task capability judgment, behavioral consistency assessment, and interest care perception, effectively avoiding the miscalibrated state of insufficient or excessive trust.

[0021] 2. Achieve transparent presentation of flight status information. Breaking through the technical bottleneck of the black box of capability information: This invention integrates three types of heterogeneous data—meteorological conditions, airspace density, and aircraft health—into a single readable index through multi-source fusion calculation of the flight readiness index. Combined with a real-time status dashboard, it presents sensor health, redundancy status, and flight envelope margin in plain language, enabling passengers to make an accurate judgment on the actual capabilities of the autonomous flight system. This solves the problem that existing cockpit interaction systems generally lack a transparent presentation mechanism.

[0022] 3. Establish a proactive and transparent disclosure mechanism for abnormal situations. Overcoming communication barriers caused by information asymmetry: This invention uses logistic regression estimation of delay probability to automatically trigger an active notification process when the delay probability exceeds a preset threshold. The reason for the delay is announced in the first person and alternative route options are pushed, conveying uncertainty to passengers in a transparent manner. This solves the shortcomings of existing solutions in the disclosure mechanism of abnormal flight status information and effectively alleviates passengers' anxiety caused by information asymmetry.

[0023] 4. Achieve adaptive control of cognitive load. Breaking through the cognitive bottleneck of information overload: This invention uses the flight phase as the information volume constraint and the passenger's working memory capacity as the upper limit benchmark to realize progressive information disclosure control based on the flight phase. The cruise phase allows the presentation of full information, while the takeoff and landing phases retain only necessary safety information. When an abnormal alarm occurs, it forces the entry into a minimized emergency mode, which solves the technical problem of existing solutions ignoring the passenger's cognitive load limit.

[0024] 5. Enhance the personalized experience of the cabin environment. Breaking through the service limitations of standardized environmental control: This invention uses a context-adaptive cabin environment control algorithm that integrates the principles of daytime and nighttime lighting with passengers' personal preferences to dynamically adjust the cabin color temperature in a weighted fusion manner. Combined with cross-flight preference memory files and a flexible takeoff emotional soothing sequence, the cabin environment control system can perceive and respond to passengers' emotional needs, enhancing passengers' perception of the system's goodwill and trust. Attached Figure Description

[0025] Figure 1 This is a schematic diagram of the overall technical solution architecture of the transparent capability trust presentation method for unmanned aerial vehicles proposed in this invention. Figure 2 This is a schematic diagram illustrating the core principle framework of the three-dimensional trust interaction framework based on the trust calibration model in this invention. Figure 3 This is a flowchart illustrating the logical flow of the multi-source fusion calculation model for the flight readiness index in this invention. Figure 4 This is a flowchart illustrating the logical flow framework of the proactive transparent disclosure mechanism for delay probability and the progressive information disclosure control based on flight phases in this invention. Figure 5 This is a schematic diagram illustrating the core principle framework of the situation-adaptive cockpit environment control algorithm in this invention; Figure 6 This is a schematic diagram illustrating the overall interaction and data flow between the service end-to-end orchestration, privacy layering architecture, and ground service personnel collaboration mechanism in this invention. Detailed Implementation

[0026] Example 1 To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to specific embodiments.

[0027] This invention provides a transparent capability trust presentation method for unmanned aerial vehicles (UAVs). The method constructs a three-dimensional trust interaction framework based on a trust calibration model. It achieves transparent presentation of capability trust through multi-source fusion calculation of flight readiness index, effectively conveys honest trust through a proactive transparent disclosure mechanism based on delay probability, and strengthens benevolent trust through progressive information disclosure control based on flight phases and a situation-adaptive cockpit environment control algorithm. The overall method includes the following core steps: Step 1: Construct a three-dimensional trust interaction framework based on a trust calibration model. This framework uses capability trust, integrity trust, and goodwill trust as its three pillars to establish a trust calibration mechanism between passengers and the autonomous flight system. The capability trust dimension centers on a real-time flight status visualization module, continuously and clearly presenting system performance data, including sensor health, redundancy status, and flight envelope margin. Information presentation uses plain language instead of technical jargon; the default mode uses a green indicator to visually indicate whether the system is operating normally, and provides real-time traffic summaries in easy-to-understand natural language. The integrity trust dimension centers on immediate route change announcements and proactive notification of abnormal conditions. When encountering route changes, delay predictions, or weather changes, the system establishes a proactive, timely, and transparent information communication process. The reasons for abnormalities are explained in the first person, ensuring that passengers' anxiety caused by information asymmetry is effectively alleviated. The Goodwill and Trust dimension is centered on cross-flight preference memory files and a gentle takeoff emotional soothing sequence. The system remembers passenger preferences and proactively adjusts the cabin environment. Soothing sound effects are activated in both ear canals during the first minute of flight. Personalized video briefings are available for viewing before departure. Augmented reality exit signs and pre-flight safety briefings in the cabin provide safety education in a non-alarming tone.

[0028] Step 2: Design a multi-source fusion calculation model for the flight readiness index. This model integrates three types of data: weather conditions, airspace density, and aircraft health. A weighted linear combination model is used to calculate the flight readiness index, and a min-max linear scaling method is used to normalize each sub-score. The final score is presented on the passenger application interface in percentage form. The weighted linear combination model for the flight readiness index uses a preset weight configuration, with the sum of the three weights being a preset total value. The system sets multiple thresholds. When the displayed flight readiness index value is greater than or equal to the first preset threshold, it is marked as ready and indicated in green. When the displayed value is between the first and second preset thresholds, it is marked as feasible and prompts the user to pay attention to weather conditions. When the displayed value is below the second preset threshold, a pre-flight manual review process is triggered, and a delay probability estimate is proactively pushed to the passenger. The weather score is calculated by integrating three factors: wind speed, visibility, and precipitation probability, using a scoring method normalized to a preset range. The airspace density score is based on real-time traffic data and uses reverse normalization; the lower the traffic density, the higher the score. The aircraft health score is output by the onboard health management system, which integrates propulsion system status, battery management system status, and structural integrity monitoring data.

[0029] Specifically, the calculation model for the Flight Readiness Index uses the following mathematical expression:

[0030] in, As a weighting factor for meteorological scores, As the spatial density score weight, Weighting of aircraft health scores The scores are weather rating, airspace density rating, and aircraft health rating, respectively, with each sub-rating normalized to the interval 0, 1. The normalization of each sub-rating uses a min-max linear scaling method, with the following formula:

[0031] The final flight readiness index is presented on the app interface in percentage form, i.e. .

[0032] Step 3: Establish a proactive and transparent disclosure mechanism for delay probability. This mechanism estimates the delay probability based on a logistic regression model, using flight readiness index, remaining time until scheduled departure, airspace traffic density, and historical delay rates for the same route as input features. When the delay probability exceeds a preset threshold, a proactive notification process is triggered, announcing the reason for the delay in a first-person perspective and pushing alternative route options. The input feature vector of the logistic regression model for delay probability includes the flight readiness index, remaining time until scheduled departure, current airspace traffic density, and historical delay rates for the same route during the same period. The model coefficients are obtained through offline training using historical operational data and are iteratively updated with new data at preset intervals. When the delay probability is greater than or equal to the preset probability threshold, the intelligent voice assistant announces the reason for the delay and the estimated waiting time in a first-person perspective, and pushes alternative route options, transparently conveying uncertainty to passengers.

[0033] The probability of delay was estimated using the following logistic regression model:

[0034] in, This represents the number of minutes remaining until the scheduled departure time. Given the current airspace traffic density, This represents the historical delay rate for this route during the same period. This is the model coefficient vector, derived offline from historical operational data. The system sets the delay probability threshold to 0.15. The active notification process is triggered at any time.

[0035] Step 4: Implement progressive information disclosure control based on flight phases. This control mechanism uses flight phases as information volume constraints and passenger working memory capacity as the upper limit benchmark. Full information is allowed during cruise, while only flight status and safety prompts are retained during takeoff and landing. When an anomaly alarm is triggered, a minimized emergency mode is forcibly entered. The calculation of passenger working memory capacity is based on information chunking theory. The number of chunks is a preset number, and the number of alternative states for each chunk is a preset number. The calculated upper limit of capacity is the preset upper limit. The load factor is dynamically assigned according to flight phases: the first preset value is used during cruise, the second preset value is used during takeoff and landing, and it is forcibly reduced to the third preset value when an anomaly alarm is triggered. The progressive information disclosure information architecture design adopts a hierarchical presentation strategy. The default mode only presents simplified states to reduce passenger cognitive load. In-depth information is available to users as needed through secondary pages or extended entry points, reserving in-depth access paths for users with proactive information needs. This strategy ensures a balance between the accessibility of safety information and passenger cognitive load control.

[0036] Specifically, the passenger's working memory capacity is calculated based on Miller's Law. An adult's working memory capacity is approximately seven information blocks, with four alternative states per block, corresponding to the typical state space of interface controls. Therefore, the maximum working memory capacity is calculated to be 14 bits. The system operates based on flight phases. As a condition, dynamically constrain the amount of information presented on the current interface. To satisfy ,in The load factor is 1.0 during the cruise phase, 0.5 during takeoff and landing, and 0.4 when an abnormal alarm is triggered.

[0037] Step 5: Develop a scenario-adaptive cockpit environment control algorithm. This algorithm calculates the target color temperature based on the local flight time. When a passenger's personal color temperature preference exists, a weighted fusion output is used to achieve a dynamic balance between daytime circadian lighting and personalized preferences. The color temperature adjustment of the full-spectrum lighting system uses a preset neutral white reference color temperature as the neutral white reference color temperature, and the adjustment range is a preset color temperature adjustment range. During the day, the high color temperature reaches a preset high color temperature threshold, and at night, the low color temperature drops to a preset low color temperature threshold. The system automatically calculates the target color temperature according to a sine function based on the local flight time. When a passenger's personal color temperature preference exists, a weighted fusion output is performed using preset fusion weights, with the passenger's historical preference as the primary factor and the daytime circadian algorithm as a secondary factor.

[0038] Specifically, the context-adaptive cockpit environment control algorithm uses color temperature (Unit: K) As the main adjustment parameter, the system automatically calculates the target color temperature based on the local flight time t (24-hour system), using the following formula:

[0039] in, The base color temperature is neutral white. To adjust the amplitude, This corresponds to the peak color temperature at noon. This is based on the passenger's personal color temperature preference stored in their file. At this time, the system replaces the pure algorithm output with weighted fusion, and the formula is:

[0040] in, The personalized weights indicate that the system is primarily driven by passengers' historical preferences, supplemented by a circadian rhythm algorithm.

[0041] A touch interaction optimization mechanism under vibration conditions. This method also includes a touch interaction optimization mechanism under vibration conditions. The cabin accelerometer samples the root mean square value of the three-axis acceleration vector in real time. When the root mean square value exceeds a preset acceleration threshold, the minimum touch target width of the safety-critical controls is automatically expanded from a first preset pixel width to a second preset pixel width, compressing the touch difficulty index to below the preset upper limit of the difficulty index, ensuring that the touch response time does not exceed the preset upper limit of the response time. Specifically, when the root mean square value of the three-axis acceleration vector exceeds the preset upper limit of the acceleration threshold, the minimum touch target width of the safety-critical controls is automatically expanded from a first preset pixel width to a second preset pixel width, compressing the touch difficulty index to below the preset upper limit of the difficulty index, ensuring that the touch response time does not exceed the preset upper limit of the response time. At the same time, the system expands the minimum touch target width of safety-critical controls from 48 dp to 72 dp, compresses the touch difficulty index to below 2.5 bits, and ensures that the touch response time does not exceed 580 ms. This mechanism quantitatively optimizes the control size and layout of the 12-inch curved OLED personal screen based on Fitts' theorem to ensure operational accessibility in any flight state.

[0042] The entire service experience is designed and orchestrated. This method also includes the design and orchestration of the entire service experience. Before flight, safety education is conducted through a personalized in-app video briefing and in-cabin augmented reality exit markers. During the first minute of flight, a gentle takeoff sequence responds to peak physiological arousal. Upon arrival, a shuttle reservation card is pushed out, and a post-trip satisfaction score is presented instantly at the peak of emotional engagement, completing a feedback loop. Specifically, the pre-flight safety briefing is designed as a 90-second personalized in-app video overlaid with a 30-second AR exit marker in the cabin, delivered in a non-alarming tone. During the first minute of flight, soothing audio effects are activated in both ears, and the intelligent voice assistant proactively explains the sensations, such as "The slight tilt you feel is normal for establishing the flight path." Upon arrival, a shuttle reservation card is pushed out before landing, and a post-trip satisfaction score is presented instantly at the peak of emotional engagement.

[0043] Privacy-Layered Architecture Design. This approach also includes a privacy-layered architecture design. Personal biometric data is stored in an encrypted passenger self-controlled file, while aircraft performance and flight path data are retained independently in a de-identified form, meeting regulatory compliance requirements while ensuring the privacy of passenger personal data. Specifically, the system establishes a two-layer data architecture. The first layer is the personal biometric data layer, which is stored in the passenger self-controlled file using AES-256 encryption. Data ownership belongs to the passenger, and access requires passenger authorization. The second layer is the operational data layer, where aircraft performance and flight path data are retained independently in a de-identified form. Passenger identification information is removed using a hash function, retaining only the data fields required for operational analysis and regulatory review.

[0044] The method also includes a collaboration mechanism between ground service personnel and the system. Ground service personnel receive real-time updates from the backend dispatch system via a dedicated operations application. The maximum response delay for operational alarms is set to a preset maximum delay, and a mandatory read receipt mechanism is implemented to ensure the timeliness of information delivery. When the intelligent system detects passenger panic, a manual takeover protocol is initiated to ensure the controllability and continuity of passenger experience during service degradation scenarios. Specifically, a WebSocket long-lived connection is established between the ground service personnel's terminal devices and the backend dispatch system to push operational alarm information in real time. The maximum response delay for operational alarms is set to 30 seconds, and the system is configured with a mandatory read receipt mechanism, with unread alarms continuously alerted with high priority. When the intelligent system collects passenger facial expression data through in-cabin cameras and detects a panic state through an emotion recognition algorithm, it automatically triggers the manual takeover protocol, notifying the nearest ground service personnel to go to the cabin for manual intervention.

[0045] The technical implementation process of the method of the present invention will be described in detail below with reference to specific application scenarios.

[0046] Application Scenario 1: Business Commute between Shenzhen and Hong Kong. Assume passenger Chen Wei has booked an eVTOL commuter flight from Shenzhen Bao'an Airport to Hong Kong International Airport for the following morning. After the booking process begins, the system uses artificial intelligence algorithms to identify the passenger's preferred route based on historical travel records. After confirming the route, the passenger selects a travel time slot, with time windows divided into 30-minute increments. After payment, the system automatically generates a QR code as a boarding pass. The confirmation page displays a real-time flight readiness index, calculated based on three data points: weather conditions, airspace density, and aircraft health. Assuming the current weather score is 0.82, airspace density score is 0.91, and aircraft health score is 0.95, the flight readiness index is... The score is 88 out of 100, and it is marked as ready in green.

[0047] 24 hours before flight departure, passengers can view a personalized safety briefing video via the passenger app. The 90-second video uses a non-alert tone to introduce the location of safety facilities and emergency exits within the cabin. Passengers can replay the video at any time within the app to reinforce their memory. 15 minutes before departure, passengers arrive at the tarmac. Ground staff use a dedicated operations app to scan the passenger's boarding pass QR code to verify their identity and guide them into the cabin. The cabin has a 2+4 six-seat layout, with the rear seats automatically adjusting to a focused, oriented mode based on the passenger's historical preferences. The system locks the seat configuration and automatically checks the seatbelt tension.

[0048] At takeoff, the cabin activates a flexible takeoff sequence. The intelligent voice assistant announces in a warm, professional tone: "You are about to begin your flight. The slight tilt you feel is normal during flight path establishment; please rest assured." Simultaneously, soothing audio in both ear canals is activated, guiding passengers' breathing rhythm with two-way sound signals. The cabin's full-spectrum RGBW lighting system automatically adjusts the color temperature based on the local flight time. For example, if the current time is 10:00 AM, the target color temperature is... It produces a bluish-white light with a color temperature close to 6000 K, typical for daytime use. If the passenger's historical preferred color temperature is 5200 K, the final blended output color temperature will be... .

[0049] During the cruise phase, the amount of information displayed on the interface is adjusted according to the load factor. Fully accessible, passengers can independently access entertainment content and preference settings. A 34-inch panoramic display shows real-time flight path and weather data, while a 12-inch personal screen allows for touch operation. The front seats are fixed and non-adjustable, while the four rear passenger seats support touchscreen reconfiguration, allowing passengers to choose between face-to-face social mode or unidirectional focused mode. The system continuously monitors cabin vibration, with accelerometers sampling the root mean square value of the three-axis acceleration vector in real time. If detected... It automatically expands the width of the touch target for safety-critical controls to 72 dp to ensure the accessibility of touch interaction.

[0050] During the flight, localized turbulence occurred, and the system determined that a minor route adjustment was needed. The route change information was immediately announced via the intelligent voice assistant in a first-person voice: "I am optimizing your flight path. Minor turbulence is expected. Rest assured, this is within my capabilities." The panoramic display screen simultaneously updated the flight path display, and the real-time status dashboard maintained a normal system operation indicator with a green light. Passenger working memory load was affected. The interface will be restricted to displaying only flight status, safety prompts, and emergency operation access points. Full information will be restored once the flight path stabilizes.

[0051] As the landing phase approaches, the system detects changes in weather conditions at the destination airport, and the delay probability estimation model is updated in real time. Assuming the current flight readiness index drops to 0.71, the remaining time δt until the scheduled departure time is negative (the scheduled departure time has passed), the current airspace traffic density ρ_a is 0.85, and the historical flight delay rate H_ph for the same period is 0.12, the following values ​​are calculated using the logistic regression model: The value exceeded the 0.15 threshold. The intelligent voice assistant proactively announced: "Due to changes in weather at the destination, the estimated arrival time will be delayed by approximately 15 minutes. I have prepared alternative connection options for you to choose from." The push interface displays alternative route options, which passengers can select via touch or have the system automatically match the optimal solution.

[0052] The flight landed smoothly on the tarmac at Hong Kong International Airport. Before the cabin door opened, the system pushed a shuttle reservation card to the passenger's mobile phone, allowing the passenger to book a shuttle vehicle to their destination in advance. At the peak of the experience, the app instantly displayed a post-flight satisfaction rating interface, inviting passengers to rate their flight experience. The rating data was imported into the passenger preference profile for future use.

[0053] Application Scenario 2: Guangzhou to Macau Leisure Travel. Assume passenger Lin Xiaotong has booked a weekend eVTOL flight from Guangzhou Baiyun Airport to Macau Heliport. This passenger has specific preferences regarding the cabin's sensory environment, has used the system multiple times, and has accumulated a personal preference profile. After the booking process is completed, the system automatically configures cabin environment parameters based on the passenger's historical preferences, including seat orientation, lighting color temperature, and music type. The flight readiness index is displayed as 92 points, marking it as ready.

[0054] During the passenger's first flight, the system activated a gentle takeoff emotional calming sequence, played a personalized safety briefing video, and displayed AR exit markers in the cabin to visually guide passengers to the emergency exits. The system recorded the passenger's sensitivity to the acoustic environment and automatically lowered the cabin background noise level on subsequent flights. The color temperature fusion output was primarily based on the passenger's preferences, supplemented by a circadian rhythm algorithm, to ensure the cabin environment met the passenger's personalized needs.

[0055] During flight, the system collects passenger voice commands via a far-field microphone array, identifies and executes requests for cabin environment adjustments. Passengers can adjust the brightness of their personal screens, select entertainment content, or check real-time flight information via gesture control or voice commands. The three-layer composite structure of the cabin walls effectively blocks external noise, keeping cabin noise levels within the range of 68 to 72 dB(A).

[0056] Upon flight arrival, the system provides recommendations for local attractions and restaurant reservations in Macau, completing a closed loop of extended travel experiences. Passenger rating data further enriches individual preference profiles, providing data support for subsequent service optimization.

[0057] Application Scenario 3: Trust Calibration Scenario under Abnormal Conditions. Assume that during flight, the onboard health management system detects an anomaly in the propulsion system parameters. The aircraft health score immediately drops, triggering a recalculation of the flight readiness index. Assuming the weather score remains at 0.80 and the airspace density score remains at 0.88, but the aircraft health score suddenly drops to 0.45, the flight readiness index will be updated to... If the score is converted to a percentage and displays as 72, which is below the first preset threshold of 85 but above the second preset threshold of 70, it is marked as "Feasible, Pay Attention to Weather" and indicated in yellow.

[0058] The delay probability estimation model is updated synchronously. Assuming the current flight readiness index is 0.715, the remaining time until the scheduled arrival time is 25 minutes, the airspace traffic density is 0.72, and the historical delay rate for the same period is 0.08, the following calculations are made: The value exceeded the 0.15 threshold. The system immediately triggered the proactive transparency disclosure mechanism, and the intelligent voice assistant announced in the first person: "I have detected a slight fluctuation in the propulsion system parameters and have initiated redundancy backup to ensure flight safety. This is not expected to affect your trip, and I am closely monitoring the status." The panoramic display screen simultaneously presented sensor health, redundancy status, and flight envelope margin data, explaining the system's current operating status in plain language.

[0059] If the aircraft health score further drops below 0.30 and the flight readiness index drops below 70, the pre-flight manual review process is triggered. The system sends an alarm to the back-end operations center, and ground service personnel receive the information via a dedicated operations app, with a response delay of no more than 30 seconds. After assessment, back-end operations personnel decide whether to execute the diversion procedure and explain the situation and decision basis to passengers via the cabin broadcast system. The passenger app simultaneously pushes delay notifications and alternative route options, transparently communicating the uncertainty to passengers.

[0060] If the intelligent system detects a passenger's panic state through its in-cabin emotion recognition algorithm, it automatically triggers a manual takeover protocol. Ground service personnel establish a voice connection with the passenger through the in-cabin intercom system, providing reassurance and real-time status updates. Ground service personnel's terminals receive real-time updates from the backend dispatch system, ensuring the timeliness and consistency of information delivery. A privacy-layered architecture ensures that passenger biometric data is used only for emotion recognition purposes, encrypted and stored in a passenger's self-controlled file, and is not mixed with operational data.

[0061] Example 2 To meet the trust-building needs of diverse user groups, this invention provides a variant embodiment of a transparent capability trust presentation method for unmanned aerial vehicles. Based on Embodiment 1, this variant embodiment adds a human service collaboration module for user groups with weaker digital habits or a highly cautious attitude towards autonomous systems.

[0062] Step S1, Manual Guidance Service Procedure. For first-time passengers or those who explicitly request manual assistance, the system automatically matches them with ground service personnel for one-on-one guidance. Ground service personnel are equipped with dedicated operation terminals, through which they receive real-time passenger information, flight status, and environmental parameters pushed by the back-end dispatch system. The operation terminals feature a large touchscreen design, with font and icon sizes optimized to accommodate users with different vision conditions. Ground service personnel scan the passenger's boarding pass QR code using the terminal to obtain passenger profile information, including historical travel records, trust profile, and special needs annotations.

[0063] Step S2, Human-AI Hybrid Service Delivery. Ground service personnel select differentiated service strategies based on passenger trust profiles. For passengers with trust anxiety regarding their capabilities, ground service personnel emphasize the aircraft's redundancy backup mechanism and sensor health monitoring system, enhancing trust perception through physical touch demonstrations. For passengers with a need for goodwill trust, ground service personnel proactively inquire about cabin environment preferences and assist in configuring personal screen parameters. For passengers lacking integrity trust, ground service personnel promise to accompany them throughout the journey and provide real-time status updates at key points.

[0064] Step S3: Service Evaluation and File Update. After the flight, ground service personnel record the service delivery details via their operational terminals, including service duration, passenger feedback, and any anomalies. This recorded data is then integrated into the passenger trust profile file for future service use. If a passenger gives a low evaluation of the service, the system automatically marks that passenger as a high-priority individual, and experienced ground service personnel will be assigned to them on the next flight.

[0065] Step S4: Privacy Protection and Data Isolation. Ground service personnel's operating terminals employ an architecture combining local encrypted storage and remote data synchronization. Passenger biometric data is stored only in passenger self-controlled files; the operating terminals do not retain sensitive data such as passenger facial recognition or fingerprints. Service record data is uploaded to the backend analysis system in de-identified form, used solely for service quality monitoring and operational optimization, and is not associated with passengers' personal identity information.

[0066] Application Scenario 4: High-priority service for traditional industry business owners. Suppose passenger Liang Guohui has booked an eVTOL flight from Hong Kong International Airport to Foshan Shadi Airport. This passenger is a 54-year-old owner of a traditional manufacturing business, with relatively weak digital habits and a highly cautious attitude towards autonomous flight systems. The system automatically identifies the passenger's trust profile, marks him as a high-priority user, and assigns him to experienced ground service personnel Chen Li for one-on-one service.

[0067] After booking confirmation, Chen Li proactively contacted the passenger through the operations terminal to confirm their travel needs and explain the service process. On the day of the flight, Chen Li arrived at the tarmac early to wait, scanned the passenger's boarding pass, and guided them into the cabin. Chen Li explained in detail the location of the cabin's safety facilities and emergency exits, and demonstrated the basic operation of the touchscreen. Addressing the passenger's concerns about flight safety, Chen Li explained the aircraft's multiple redundancy backup mechanisms and real-time health monitoring system in simple terms, and displayed historical data records from the real-time status dashboard.

[0068] Before takeoff, Chen Li confirmed that passengers' seatbelts were properly fastened and their seats were locked. At takeoff, Chen Li provided real-time voice narration via the in-flight intercom system: "The aircraft is taking off. You will feel a slight ascent, which is normal, please rest assured." During cruise, Chen Li periodically updated the flight status via the in-flight intercom system to alleviate passenger anxiety. If a route adjustment was detected, Chen Li immediately explained the reason and corresponding countermeasures via voice.

[0069] As the landing approached, Chen Li informed passengers of the arrival time and shuttle vehicle information in advance. After the cabin door opened, Chen Li guided passengers safely off the plane and assisted them to the shuttle area. After the service, Chen Li recorded the service details through the operations terminal, and the passengers rated the service as satisfactory. This evaluation data was entered into the passenger's file, and the system will automatically match a service personnel of the same level for the next flight.

[0070] The foregoing has shown and described the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The embodiments and descriptions in the specification are merely illustrative of the principles of the invention. Various changes and modifications can be made to the invention without departing from its spirit and scope, and all such changes and modifications fall within the scope of the present invention as claimed. The scope of protection of this invention is defined by the appended claims and their equivalents.

Claims

1. A method for presenting transparent capability trust for unmanned aerial vehicles, characterized in that, Includes the following steps: Step 1: Construct a three-dimensional trust interaction framework based on a trust calibration model: using capability trust, integrity trust, and goodwill trust as the three pillars of the design, establish a trust calibration mechanism between passengers and the autonomous flight system; the capability trust dimension is realized through a real-time flight status visualization module, with the system continuously presenting system performance data; the integrity trust dimension is realized through an interactive process that proactively informs passengers of abnormal states; and the goodwill trust dimension is realized through a personalized context perception module. Step 2: Design a multi-source fusion calculation model for the flight readiness index: Integrate three types of data, namely meteorological conditions, airspace density and aircraft health, and use a weighted linear combination model to calculate the flight readiness index. Use a normalization method to normalize each sub-score and finally present it in the passenger application interface in the form of a percentage. Step 3: Establish a proactive and transparent disclosure mechanism for delay probability: Estimate the delay probability based on the prediction model, and take the flight readiness index, the remaining time until the scheduled departure time, the airspace traffic density and the historical flight delay rate of the same period as input features. When the delay probability exceeds the preset threshold, a proactive notification process is triggered to announce the cause of the delay in the first person and push alternative route options. Step 4, implement progressive information disclosure control based on flight phase: use flight phase as the information volume constraint and passenger working memory capacity as the upper limit benchmark. During the cruise phase, full information is allowed to be presented. During takeoff and landing phases, only flight status and safety prompts are retained. When an abnormal alarm is triggered, the system is forced to enter the minimized emergency mode. Step 5: Develop a scenario-adaptive cockpit environment control algorithm: Calculate the target color temperature based on the local flight time, and use weighted fusion output when there are passengers' personal color temperature preferences to achieve a dynamic balance between daytime circadian lighting and personalized preferences.

2. The method for presenting transparent capabilities and trust in unmanned aerial vehicles according to claim 1, characterized in that: The design of the capability trust dimension in step 1 is centered on the real-time flight status visualization module. The system continuously and clearly presents system performance data, including sensor health, redundancy status, and flight envelope margin. Information is presented in plain language instead of technical terms, and the default mode uses visual indicators to intuitively show whether the system is operating normally.

3. The method for presenting transparent capabilities and trust in unmanned aerial vehicles according to claim 1, characterized in that: The design of the integrity and trust dimension in step 1 is centered on real-time announcements of route changes and proactive notification of abnormal situations. When encountering route changes, delay predictions, or weather changes, the system establishes a proactive, timely, and transparent information communication process. The reasons for the abnormality are explained in the first person to ensure that passengers' anxiety caused by information asymmetry is effectively alleviated.

4. The method for presenting transparent capabilities and trust in unmanned aerial vehicles according to claim 1, characterized in that: The design of the benevolent trust dimension in step 1 is based on the cross-flight preference memory archive and the gentle takeoff emotional soothing sequence. The system remembers passenger preferences and actively adjusts the cabin environment. The soothing sound effect in both ear canals is activated in the first minute of the flight. The personalized video briefing in the application can be viewed before departure. The augmented reality exit markings in the cabin and the pre-flight safety briefing are completed in a non-alarming tone to provide safety education.

5. The method for presenting transparent capabilities and trust in unmanned aerial vehicles according to claim 1, characterized in that: In step 2, the weighted linear combination model of the flight readiness index adopts a preset weight configuration, and the sum of the three weights is a preset total value. The system sets multiple thresholds. When the flight readiness index is greater than or equal to the first preset threshold, it is marked as ready and indicated in green. When the displayed value is between the first and second preset thresholds, it is marked as feasible and prompts passengers to pay attention to the weather. When the displayed value is lower than the second preset threshold, the pre-flight manual review process is triggered and the system proactively pushes a delay probability estimate to passengers.

6. The method for presenting transparent capabilities and trust in unmanned aerial vehicles according to claim 1, characterized in that: In step 2, the meteorological score is calculated by taking into account three factors: wind speed, visibility, and precipitation probability, and a scoring method that is normalized to a preset interval is used. The airspace density score is based on real-time traffic data and uses reverse normalization processing; the lower the traffic density, the higher the score. The aircraft health score is output by the airborne health management system, which integrates propulsion system status, battery management system status, and structural integrity monitoring data.

7. The method for presenting transparent capabilities and trust in unmanned aerial vehicles according to claim 1, characterized in that: In step 3, the input feature vector of the delay probability prediction model includes the flight readiness index, the remaining time until the scheduled takeoff time, the current airspace traffic density, and the historical delay rate of the same route during the same period. The model coefficients are obtained by offline training from historical operational data and are iteratively updated with new data at a preset period. When the delay probability is greater than or equal to the preset probability threshold, the intelligent voice assistant announces the cause of the delay and the estimated waiting time in the first person and pushes alternative route options.

8. The method for presenting transparent capability trust for unmanned aerial vehicles according to claim 1, characterized in that: In step 4, the calculation of passenger working memory capacity is based on the information chunking theory. The number of chunks is a preset number, and the number of alternative states for each chunk is a preset number. The calculated upper limit of capacity is the preset upper limit of capacity. The load factor is dynamically assigned according to the flight phase. The first preset value is used during the cruise phase, the second preset value is used during the takeoff and landing phases, and the value is forcibly reduced to the third preset value when an abnormal alarm is triggered. The information architecture design of progressive information disclosure adopts a hierarchical presentation strategy. The default mode only presents simplified states to reduce the cognitive load on passengers, and in-depth information is available for users to access as needed through secondary pages or extended entry points.

9. The method for presenting transparent capability trust for unmanned aerial vehicles according to claim 1, characterized in that: In step 5, the color temperature adjustment of the full-spectrum lighting system uses a preset neutral white reference color temperature as the neutral white reference color temperature, and the adjustment range is a preset color temperature adjustment range. During the day, the high color temperature reaches the preset high color temperature threshold, and at night, the low color temperature drops to the preset low color temperature threshold. The system automatically calculates the target color temperature according to the local flight time and a preset function rule. When there is a passenger's personal color temperature preference, a preset fusion weight is used for weighted fusion output, with the passenger's historical preference as the main factor and the daytime rhythm algorithm as the auxiliary factor.

10. The method for presenting transparent capability trust for unmanned aerial vehicles according to claim 1, characterized in that: It also includes a touch interaction optimization mechanism for vibration environments. The cabin acceleration sensor samples the root mean square value of the three-axis acceleration vector in real time. When the root mean square value exceeds the preset acceleration threshold, the minimum touch target width of the safety-critical control is automatically expanded from the first preset pixel width to the second preset pixel width, compressing the touch difficulty index to below the preset difficulty index upper limit, ensuring that the touch response time does not exceed the preset response time upper limit. It also includes a full-process service experience orchestration design. In the pre-flight stage, safety education is completed through personalized video briefings in the application and augmented reality exit labels in the cabin. In the first minute of flight, a flexible takeoff sequence is used to respond to the peak of physiological arousal. In the arrival stage, a connection reservation card is pushed and a post-trip satisfaction score is presented in real time when emotions are at their peak. It also includes a privacy layered architecture design. Personal biometric data is stored in an encrypted passenger self-control file, and aircraft performance and flight path data are retained independently in a de-identified form. It also includes a collaboration mechanism between ground service personnel and the system. Ground service personnel receive backend scheduling updates in real time through a dedicated operation application. The maximum response delay for operation alarms is set to the preset maximum response delay. Forced read receipts are set to ensure the timeliness of information transmission. When the intelligent system detects passenger panic, a manual takeover protocol is initiated.

Images