Method for managing motion sickness and system for implementing it
A method using multiple algorithms and a learning algorithm to accurately assess motion sickness intensity and execute tailored countermeasures addresses the issue of unreliable motion sickness detection, improving estimation reliability.
Patent Information
- Authority / Receiving Office
- FR · FR
- Patent Type
- Applications
- Current Assignee / Owner
- AMPERE SAS
- Filing Date
- 2024-12-16
- Publication Date
- 2026-06-19
AI Technical Summary
Existing systems fail to provide an accurate estimation of motion sickness intensity, leading to false positives or negatives, necessitating improved reliability in motion sickness detection.
A method using multiple algorithms to determine theoretical motion sickness intensity based on vehicle and occupant parameters, combined with a probabilistic approach and a learning algorithm to refine these estimates, followed by a control mechanism to execute appropriate countermeasures.
Enhances the accuracy of motion sickness estimation by optimizing countermeasures based on probabilistic analysis, reducing false positives and negatives.
Smart Images
Figure 00000015_0000 
Figure 00000015_0001
Abstract
Description
Title of the invention: Method for managing motion sickness and system for implementing it. Technical field
[0001] The present invention relates, in general, to the management of motion sickness in vehicles.
[0002] More specifically, the invention relates to a method for managing motion sickness and to a system for implementing this management method in a motor vehicle. Previous techniques
[0003] Research on the detection of motion sickness in motor vehicles or driving simulators is well documented and many algorithms are proposed, using physiological data from the vehicle occupants, sometimes combined with data related to the movement of the vehicle.
[0004] For example, from international application WO 2020 / 193127, a device for detecting motion sickness in a person in a vehicle is known, configured to generate a motion sickness index value from the electrodermal response and the person's skin temperature.
[0005] In general, however, no system can provide an estimate of the intensity of motion sickness that is accurate enough to avoid cases of false positives or false negatives.
[0006] There therefore remains a need to improve the reliability of the estimation of the intensity of motion sickness. Description of the invention
[0007] The invention aims in particular to overcome this drawback.
[0008] The invention proposes a method for managing motion sickness during a journey in a vehicle, comprising the steps of:
[0009] - provide a plurality of distinct algorithms each configured to determine a theoretical intensity of motion sickness for a vehicle occupant as a function of at least one predetermined parameter;
[0010] - collect data representative of said predetermined parameters;
[0011] - process said collected data with said plurality of algorithms in order to deduce with each algorithm a theoretical motion sickness intensity and a statistical robustness associated with said theoretical motion sickness intensity;
[0012] - analyze said theoretical motion sickness intensities and said robustness statistics to deduce the most likely intensity of motion sickness for the vehicle occupant as well as the probability associated with this most likely intensity;
[0013] - deduce from said most probable intensity at least one measure to be taken against the motion sickness; and
[0014] - to control the execution of said measure to be taken according to said probability associate.
[0015] In other words, the method according to the invention uses a probabilistic approach to assess the intensity of motion sickness. By having the intensity of motion sickness as well as the probability associated with this intensity, one optimizes not only the measure to be taken against motion sickness but also the control of the triggering of this measure.
[0016] According to other features, the step of deducing a measure to be taken against motion sickness includes a step of identifying, among a set of measures against motion sickness, each associated with a distinct range of motion sickness intensities, the one whose associated range of motion sickness intensities includes said most probable intensity, and in which the step of controlling the execution of said measure to be taken includes a step of comparing said probability associated with a trigger threshold.
[0017] According to other features, the step of analyzing said theoretical motion sickness intensities and said statistical robustness includes the use of a learning algorithm configured to weight said theoretical motion sickness intensities, said learning algorithm being trained on training data including at least one real motion sickness intensity reported by an occupant of said vehicle during a previous journey.
[0018] In addition, the step of analyzing said theoretical motion sickness intensities and said statistical robustness includes the step of performing a weighted average of said theoretical motion sickness intensities and / or said statistical robustness.
[0019] Advantageously, the learning algorithm is a neural network.
[0020] According to other features, at least one of said algorithms is configured to determine a theoretical motion sickness intensity as a function of a parameter related to the vehicle, and at least one of said algorithms is configured to determine a theoretical motion sickness intensity as a function of a physiological parameter of the vehicle occupant.
[0021] According to other features, at least one said algorithm is configured to determine a theoretical motion sickness intensity as a function of the occupant's position in the vehicle.
[0022] According to other features, the step of processing said data collected with said plurality of algorithms is repeated following a variable time step determined according to the most probable intensity of motion sickness, the associated probability and / or their evolution over time.
[0023] According to other features, the step of processing said data collected with said plurality of algorithms is repeated following a variable time step determined according to the driving context and / or its evolution during the journey made by the vehicle.
[0024] The invention further relates to a motion sickness management system for a vehicle, comprising sensors configured to provide data representative of predetermined parameters, a human-machine interface configured to provide at least one actual motion sickness intensity declared by an occupant of said vehicle and a controller configured to collect from said sensors and said human-machine interface said data representative of predetermined parameters and said actual motion sickness intensity to implement the steps of the process as defined above. Brief description of the drawings
[0025] Other purposes, advantages and features will become apparent from the following description, given for illustrative purposes only and with reference to the accompanying drawings on which:
[0026] Figure 1 illustrates the structure of a system configured to implement a motion sickness management method according to the invention; and
[0027] [Fig.2] is a diagram illustrating the steps in the motion sickness management process. DETAILED DESCRIPTION
[0028] Fig. 1 illustrates the structure of a motion sickness management system 10 intended to be fitted to a motor vehicle (not shown).
[0029] The system 10 includes a sensor 11 configured to provide vehicle motion data, a sensor 12 configured to provide physiological data of a vehicle occupant, a human-machine interface 13, a controller 14 and a communication bus 15 to which the sensors 11 and 12, the interface 13 and the controller 14 are connected.
[0030] Sensor 11 is, for example, an acceleration sensor measuring acceleration along the forward-backward, left-right and up-down directions conventionally defined with respect to the vehicle.
[0031] Sensor 12 is, for example, an electrodermal response sensor, measuring the electrical conductance of the skin. The electrodermal response is a parameter physiological allowing to evaluate the intensity of perspiration on the surface of an individual's skin, the intensity of perspiration being known to be correlated with that of motion sickness that this individual might experience.
[0032] The human-machine interface 13 here includes an interactive display device 16, for example a touch screen, and a speaker 17.
[0033] The controller 14, for example a vehicle computer, includes a data processing unit 18 and a memory 19 connected to the processing unit 18.
[0034] The communication bus 15 is for example the CAN bus (for Controller Area Network in English) of the vehicle.
[0035] The controller 14 is configured to collect the data provided by the sensors 11 and 12 as well as the interactive display device 16, and to record this data in the memory 19.
[0036] The controller 14 further includes a set of motion sickness management modules during a journey with the vehicle, comprising a first module 30, a second module 31 and a third module 32, here each recorded in memory 19.
[0037] Module 30 is configured to determine the most probable intensity of motion sickness iprob for the vehicle occupant as well as the probability P(iprob) associated with this intensity iprob.
[0038] Module 30 includes two separate algorithms 20 and 21 for determining theoretical motion sickness intensities, an analysis algorithm 22 for performing a probability analysis of the results provided by algorithms 20 and 21, and a learning algorithm 23 for customizing the analysis algorithm 22 to the vehicle occupant.
[0039] Algorithm 20 is configured to process the representative data of vehicle acceleration collected by sensor 11 to deduce a first value of theoretical motion sickness intensity ithi as well as a first statistical robustness ri associated with this first value of theoretical motion sickness intensity ithl.
[0040] Algorithm 20 is based here on the relationships:
[0041] hfti — a ' ttrms + fi
[0042] y '1 - «HW
[0043] where a and fi are coefficients obtained by learning, arms is the root mean square of the components ax(t), ay(t), and az(t) of the acceleration measured by sensor 11 along the front-back, left-right and up-down directions respectively, for a duration T, namely:
[0044]
[0045] and is the standard deviation of the acceleration values.
[0046] For example, if arms = 2.5, mis 2, a = 0.8, fl = 1 and <7ams = 0.5, then:
[0047] = 0.8-2.5+1 = 3.0
[0048] r, = g=0.2
[0049] Algorithm 21 is configured to process representative electrodermal response data collected by sensor 12 to deduce a second theoretical motion sickness intensity value ith2 and a second statistical robustness r2 associated with this second theoretical motion sickness intensity value ith2.
[0050] Algorithm 21 is based here on the relationships:
[0051] ith2 = y • max (Gfast) + 5 • Gslow
[0052] _ r2 ~ max(Gfap
[0053] where and <5 are coefficients obtained by learning, Gfast(t) and Gslow(t) are respectively the fast variation and the slow variation extracted from the same raw signal GSR(t) provided by the sensor 12 (GSR being the acronym for the English expression Galvanic Skin Responsey std a function returning the standard deviation etmax a function returning a maximum.
[0054] For example, if max(Gfast) — 3.5 / iS, Gslow = 1.2pS* y = 0.7, 5 = 0.3 and Std ( Gf^ ) = 0.4, then:
[0055] ith2 = 0.7 • 3.5 + 0.3-1.2 = 2.45 + 0.36 = 2.81
[0056] r2 = |4=0114
[0057] It should be noted that algorithms 20 and 21 are distinguished here by the measured parameter from which they each determine the theoretical intensity of motion sickness, namely here the acceleration of the vehicle and the electrodermal response of the vehicle occupant.
[0058] It should be noted that algorithms 20 and 21 are further distinguished here by the relationship, empirical in this case, that they each establish between the measured parameter and the corresponding theoretical motion sickness intensity ithi or ith2.
[0059] Alternatively, algorithms 20 and 21 could differ only in the relationship they each establish between the measured parameter and the corresponding theoretical motion sickness intensity, the measured parameter being the same for each algorithm.
[0060] The analysis algorithm 22 is configured to perform a probability analysis step of the theoretical motion sickness intensity values ithi, ith2 as well as the statistical robustness values rb r2 determined by algorithms 20 and 21, in order to deduce the most probable intensity of motion sickness iprob as well as the associated probability P(iprob).
[0061] During the probability analysis step, weighting coefficients ki and k2 are assigned to a theoretical motion sickness intensity ithb ith2 respectively.
[0062] Algorithm 22 is based here on the relationships:
[0063] • _ Iprob ~ k^k2
[0064] where ki = 1 / and h — * / , and
[0065] p / • x ,
[0066] In other words, the probability analysis step includes determining the most probable motion sickness intensity iprob as being equal to the weighted average of the theoretical intensity values ithi, ith2 using the weighting coefficients kh k2.
[0067] The associated probability P(iprob) is, in this case, a probability of similarity of the theoretical intensity values ithi, ith2.
[0068] For example, with the values calculated above for ithi, rb ith2 and r2, we obtain:
[0069] ^ = ^=5
[0070] ^ = _!_œ8.77 [°°71i [°°721 PM = 1-^ = 1-^=0.937
[0073] The learning algorithm 23, here a neural network, is configured to adjust the weighting coefficients kb k2 according to training data collected during at least one prior journey made by the occupant with the vehicle.
[0074] The training data includes, for each previous journey, at least one most probable motion sickness intensity iprob, the data from sensors 11 and 12 from which this intensity iprob was determined, and at least one actual motion sickness intensity idec reported by the vehicle occupant. The actual motion sickness intensity idec is, for example, reported using the interactive display device 16, or a voice-activated device.
[0075] The learning algorithm 23 is configured here to implement backpropagation adjustment to minimize the value |ip^ - i^ec| by adjusting the weighting coefficients ki and k2. The learning algorithm 23 is based on the loss function:
[0076] L = (îprob~idec) 2
[0077] The learning algorithm 23 adjusts the coefficients ki and k2 to minimize this loss, refining the correspondence between the calculated motion sickness intensities and the reported motion sickness intensities.
[0078] Module 31 is configured to deduce from the most probable intensity iprob a suitable measure against motion sickness and to provide instructions for implementing this measure to controller 14.
[0079] Module 31 includes for this purpose a plurality of instruction blocks, here five blocks 24, 25, 26, 27 and 28, as well as a selection algorithm 29 configured to identify an instruction block 24 to 28 to be executed according to the most probable intensity iprob.
[0080] Each instruction block 24 to 28 is associated with a range of values pvib24, pvib25, pvib26, pvib27 and pvib28 of the intensity of motion sickness and allows the controller 14 to implement a measure against motion sickness adapted to the corresponding range of intensity values.
[0081] The ranges of values pvib24 to pvib28 are for example respectively [0; 2[, [2; 4[, [4; 6[, [6; 8[ and [8; 10], the values 0, 2, 4, 6, 8 and 10 corresponding respectively to an intensity of motion sickness of zero, low, medium, high, very high and maximum on the discomfort scale from 0 to 10.
[0082] The selection algorithm 29 is configured to find among the instruction blocks 24 to 28 the one whose associated range of values pvib24 to pvib28 includes the motion sickness intensity value iprob, and to generate an identifier for this instruction block.
[0083] For example, with the value iprob = 2.89 calculated above, algorithm 29 will select block of instructions 25.
[0084] Module 32 is configured to control the execution of the block of instructions to be executed according to the associated probability P(iprob).
[0085] Module 32 includes for this purpose a triggering algorithm 33 configured to compare the probability P(iprob) with a triggering threshold sded and generate a binary triggering control value, for example, equal to 1 if the probability P(iprob) is greater than or equal to the triggering threshold sdeci, and equal to 0 otherwise. The triggering threshold sded is, for example, equal to 0.8. Given the value of P(iprob) = 0.937 calculated above, the triggering algorithm 33 will set the binary triggering control value to 1.
[0086] The controller 14 is configured to execute the block of instructions identified by module 31 if the control value generated by module 32 is equal to 1.
[0087] Fig. 2 is a diagram illustrating steps of a motion sickness management process implemented by system 10 during a journey with the vehicle.
[0088] During a step El, the algorithms 20 and 21 are provided by being stored in the memory 19 of the controller 14. The algorithms 20 and 21 are for example found in technical literature available on an online database, such as a patent database.
[0089] During a step E2, the controller 14 collects representative data of vehicle acceleration and of the electrodermal response of the vehicle occupant provided by sensors 11 and 12 respectively, and records them in memory 19.
[0090] During a step E3, the data processing unit 18 processes the data obtained in step E2 with each of the algorithms 20 and 21 to deduce the first theoretical motion sickness intensity value ithb, the first statistical robustness rb, the second theoretical motion sickness intensity value ith2 and the second statistical robustness r2, the values ithb, rb, ith2 and r2 being stored in memory 19.
[0091] Step E3 is repeated here following a fixed time step, on the order of 3 minutes, the values ithb rb ith2 and r2 recorded in memory 19 being updated at each iteration.
[0092] During a step E4, the data processing unit 18 uses the learning algorithm 23 to assign the coefficients ki and k2 to the values ithb ith2 respectively and / or to the values rb r2 respectively and accordingly updates the values ithb rb ith2 and r2 stored in memory 19.
[0093] During a step E5, the data processing unit 18 uses the analysis algorithm 22 to perform a probability analysis with the values ithb rb ith2 and r2 found in memory 19 and deduces the most probable value of motion sickness intensity iprob for the vehicle occupant as well as the associated probability P(iprob), then stores the values iprob and P(iprob) in memory 19.
[0094] During a step E6, the data processing unit 18 uses the selection algorithm 29 to find, among the instruction blocks 24 to 28, the one whose associated range of values pvib24 to pvib28 includes the motion sickness intensity value iprob, and generates an identifier for this instruction block. The data processing unit 18 then stores the identifier in memory 19.
[0095] It should be noted that this step E6 amounts in practice to deducing from the most probable intensity of motion sickness, defined by its iprob value, a measure to be taken against motion sickness, defined by its corresponding block of instructions.
[0096] During a step E7, the data processing unit 18 uses the triggering algorithm 33 to compare the probability P(iprob) with the triggering threshold Sdeci and generates a binary triggering control value, for example equal to 1 if the probability P(iprob) is greater than or equal to the trigger threshold sded, and equal to 0 otherwise. The data processing unit 18 stores the trigger control value in memory 19.
[0097] During a step E8, the controller 14 consults the memory 19 to find the instruction block identifier and the binary control value and executes the instructions contained in the instruction block if the binary control value is equal to 1.
[0098] It should be noted that steps E7 and E8 amount in practice to checking the execution of the measure to be taken against motion sickness identified in step E6 as a function of the associated probability P(iprob).
[0099] The motion sickness measure associated with instruction block 24 (motion sickness intensity from 0 to 2) here includes the action of stopping another motion sickness measure possibly in progress.
[0100] The motion sickness measure associated with instruction block 25 (motion sickness intensity of 3 or 4) is a preventive motion sickness measure, including for example the action of emitting audible instructions with the loudspeaker 17 in order to anticipate driving contexts that may aggravate motion sickness, such as a winding road.
[0101] The motion sickness measure associated with instruction block 26 (motion sickness intensity of 5 or 6) is here a corrective motion sickness measure, including for example the emission of a binaural sound with the loudspeaker 17, the lowering of the temperature in the vehicle's passenger compartment and / or the activation of a ventilation.
[0102] In the illustrated embodiment, step E3 is repeated at fixed intervals. Alternatively, the time interval is variable to optimize the computing power of controller 14. The variable time interval can be determined based on the most probable intensity of motion sickness iprob, the associated probability P(iprob), and / or the time evolution of these values iprob and P(iprob). For example, if the intensity iprob increases with each iteration while the probability P(iprob) remains high, then the time interval will be decreased, for example, to 1 minute. Conversely, the time interval can be increased, for example, to 10 minutes, if the probability P(iprob) remains low for several iterations. Alternatively, the variable time interval is determined based on data representative of the driving context and / or its evolution during the vehicle's journey.Such data, representative of the driving context, includes, for example, the type of road (city road, country road, motorway, etc.), the outside temperature, the temperature inside the passenger compartment, the altitude, etc. Alternatively, the pitch of... Variable time is determined based on the elapsed travel time, for example, it decreases as the elapsed travel time increases.
[0103] In one embodiment, the training data may further include an identifier specific to the vehicle occupant who declared the intensity idec. Thus, the learning algorithm 23 can be configured to distinguish the training data specific to one vehicle occupant from that specific to another vehicle occupant and generate weighting coefficients kb k2 specific to each occupant.
[0104] In one embodiment, the management process does not include the step of using a learning algorithm.
[0105] In one embodiment, step E6 of deducing a measure to be taken against motion sickness from the most probable intensity of motion sickness is only carried out if the probability P(iprob) is greater than the triggering threshold sded.
[0106] In one embodiment, module 31 comprises more or less than five instruction blocks, for example three blocks, or seven blocks.
[0107] In one embodiment, algorithm 20 and / or algorithm 21 is further configured to determine the theoretical intensity of motion sickness based on the occupant's position in the vehicle. For example, algorithm 20 and / or algorithm 21 applies a surcharge coefficient if the occupant's position is a passenger seat.
[0108] In one embodiment, the first module 30 comprises more than two distinct algorithms, such as 20 and 21, for determining a theoretical motion sickness intensity, for example 3, 10 or 100 algorithms.
[0109] In one embodiment, at least one of the algorithms such as 20 or 21 is configured to determine a theoretical motion sickness intensity for a vehicle occupant as a function of a plurality of predetermined parameters, the predetermined parameters being selected from: - physiological parameters of the vehicle occupant, such as body temperature, skin temperature, electrodermal response, respiratory rate, exhaled oxygen concentration, exhaled oxygen concentration, heart rate and / or heart rhythm - a parameter identifying the occupant's position in the vehicle, such as the driver's seat or a passenger seat located at the front, rear, left, right or middle; - vehicle movement parameters, such as speed, acceleration and / or vibration; - atmospheric parameters, such as air temperature, oxygen level, oxygen anion level, dioxygen level and / or noise level; - geographical parameters, such as altitude and / or GPS coordinates; - other parameters such as the type of road taken by the vehicle (city road, country road, motorway etc.), the outside temperature and / or the temperature inside the passenger compartment; the system then includes, in place of sensors 11 and 12 or in addition to sensors 11 and 12, other sensors configured to collect data representative of these predetermined parameters, for example sensor 11 is replaced or supplemented by speed, vibration or steering wheel angle sensors, while sensor 12 is replaced or supplemented by body temperature, heart rate, oxygen and / or CO2 concentration sensors in exhaled air.
[0110] It should be noted that vehicle movement parameters are more generally part of vehicle-related parameters, the latter also including, for example, engine speed, steering wheel angle, seat back angle, ABS system on or off, air conditioning setpoint temperature, cabin temperature, open or closed window status, etc.
[0111] In one embodiment, the vehicle equipped with the motion sickness management system is selected from among a land vehicle, an aircraft and a ship, the vehicle being real or virtual.
Claims
Demands
1. A method for managing motion sickness during a journey in a vehicle, characterized in that it comprises the steps of: - providing (E1) a plurality of distinct algorithms (20, 21), each configured to determine a theoretical motion sickness intensity for a vehicle occupant as a function of at least one predetermined parameter; - collecting (E2) data representative of said predetermined parameters; - processing (E3) said collected data with said plurality of algorithms to deduce, with each algorithm (20, 21), a theoretical motion sickness intensity and a statistical robustness associated with said theoretical motion sickness intensity; - analyzing (E5) said theoretical motion sickness intensities and said statistical robustness to deduce a most probable motion sickness intensity for the vehicle occupant as well as the probability associated with this most probable intensity;- deduce (E6) from said most probable intensity at least one measure to be taken against motion sickness; and - monitor (E7, E8) the execution of said measure to be taken according to said associated probability.
2. A method according to claim 1, wherein the step of deducing (E6) a measure to be taken against motion sickness comprises a step of identifying, among a set of measures against motion sickness, each associated with a distinct range of motion sickness intensities, the one whose associated range of motion sickness intensities includes said most probable intensity, and wherein the step of controlling the execution of said measure to be taken comprises a step of comparing said probability associated with a trigger threshold.
3. A method according to claim 1 or 2, wherein the step of analyzing (E5) said theoretical motion sickness intensities and said statistical robustness comprises the use (E4) of a learning algorithm (23) configured to weight said theoretical motion sickness intensities, said learning algorithm (23) being trained on training data comprising at least an actual intensity of motion sickness reported by an occupant of said vehicle during a previous journey.
4. A method according to claim 3, wherein the step of analyzing (E5) said theoretical motion sickness intensities and said statistical robustness comprises the step of performing a weighted average of said theoretical motion sickness intensities.
5. A method according to claim 3 or 4, wherein the learning algorithm (23) is a neural network.
6. A method according to any one of claims 1 to 5, wherein at least one of said algorithms (20) is configured to determine a theoretical motion sickness intensity as a function of a vehicle-related parameter, and at least one of said algorithms (21) is configured to determine a theoretical motion sickness intensity as a function of a vehicle occupant's physiological parameter.
7. A method according to any one of claims 1 to 6, wherein at least one of said algorithms is configured to determine a theoretical motion sickness intensity as a function of the occupant's position in the vehicle.
8. A method according to any one of claims 1 to 7, wherein the step of processing (E3) said data collected with said plurality of algorithms (20, 21) is repeated at a variable time step determined according to the most probable intensity of motion sickness, the associated probability and / or their evolution over time.
9. A method according to any one of claims 1 to 8, wherein the step (E3) of processing said collected data with said plurality of algorithms (20, 21) is repeated at a variable time step determined according to the driving context and / or its evolution during the journey made by the vehicle.
10. A motion sickness management system (10) for a vehicle, comprising sensors (11, 12) configured to provide data representative of predetermined parameters, a human-machine interface (13) configured to provide at least one actual motion sickness intensity reported by an occupant of said vehicle, and a controller (14) configured to collect from said sensors (11, 12) and said human-machine interface (13) said data representative of predetermined parameters and said actual motion sickness intensity to implement the steps of the process according to any one of claims 4 to 9, each according to claim 3.