CN102707305B - Tumble detecting and positioning system and method - Google Patents
Tumble detecting and positioning system and method Download PDFInfo
- Publication number
- CN102707305B CN102707305B CN201210185040.9A CN201210185040A CN102707305B CN 102707305 B CN102707305 B CN 102707305B CN 201210185040 A CN201210185040 A CN 201210185040A CN 102707305 B CN102707305 B CN 102707305B
- Authority
- CN
- China
- Prior art keywords
- module
- information
- accelerometer
- network
- pitch
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Landscapes
- Alarm Systems (AREA)
- Navigation (AREA)
Abstract
The invention discloses a tumble detecting and positioning system and method. The tumble detecting and positioning system comprises a tumble detecting and positioning device, which comprises an inertia navigation module, a GPS (Global Position System) module, a main controller and an Sim (Subscriber Identity Module)-300 module, wherein the inertia navigation module consists of a triaxial gyroscope, a triaxial magnetometer and a triaxial accelerometer; the GPS module and the inertia navigation module are respectively connected with the main controller; the main controller is connected with the Sim-300 module; and the tumble detecting and positioning device is connected with a kin mobile phone and a ward station, which are bound with tested personnel, through GSM (Global System for Mobile Communications)/GPRS (General Packet Radio Service). A tumble detection algorithm based on a nerve network and machine learning is further designed to accurately detect the tumble status and position information of human bodies, the information is transmitted to the ward station through the GSM/GPRS, and the status and the positions of tested personnel are real-timely displayed on a monitoring picture. The tumble detecting and positioning system can accurately detect the tumble status of the tested personnel and track and position the tested personnel in real time, is convenient and practical, and has high accuracy and strong stability.
Description
Technical field
The present invention relates to a kind of behavioural analysis and recognition methods, condition monitoring and warning, detection and positioning system, relate in particular to one and fall down detection and location system.
Background technology
21 century is the epoch of aging population, and at present, all developed countries have all entered aging society in the world, and many developing countries or be about to enter aging society.China is the maximum country of elderly population in the world, show according to data, in the end of the year 2004, China 60 years old and above elderly population are 1.43 hundred million, within 2014, will reach 200,000,000, within 2026, will reach 300,000,000, accidental falls is the main health threat of over-65s crowd and the cause of death, and because falling down in the crowd who needs medical treatment and nursing, 65 years old and above accounting for exceed 30%, and because falling down in lethal crowd, 40% is 80 years old above the old.In the crowd who exceedes 85 years old, 2/3 accidental falls directly causes death.It was reported demonstration, in the family and home for destitute in old age, 66% resident accidental falls at least once every year, and this does not comprise the not situation of report, therefore, this numeral or may be underestimated.The elderly falls down that incidence is high, consequence is serious, has become a serious medical care problem and social concern, effectively predicts that by the means of science the elderly falls down, and has become the new study hotspot in the whole world thereby reduce the elderly because falling down the injury bringing.
At present, detect for falling over of human body, from hardware, be generally divided into based on vision with based on two kinds of Wearable sensors, wherein, adopt vision to carry out falling over of human body detection and can seriously be subject to external environment impact, such as illumination condition, background, block size and video camera quality etc., in addition, because video camera monitored area is limited, monitored the elderly or patient's scope of activities can be restricted, in the research that utilizes Wearable sensor human body to fall down, a kind of is the acceleration that adopts the activity of accelerometer human body, judge whether to fall down by setting threshold, this method is difficult to distinguish falls down the aggravating activities daily with people, as jumped, above go downstairs etc.Patent 200720125141.1, 200910145045.7 be all to adopt a three axis accelerometer to record human body acceleration, calculated angle of inclination simultaneously, the former judges whether to fall down by setting acceleration and angle threshold, be difficult to distinguish the vigorous motions such as quick walking and stair activity, the latter judges that people is subject to impacting the angular relationship of front and back a period of time in the process of falling down and judges whether to fall down, the method requires falling over of human body process to occur obviously impacting, being difficult to identify the elderly falls in a swoon suddenly or falls down by a small margin, and, its angle is to obtain by acceleration calculation, obviously, the angle of inclination of calculating in the time of human body aggravating activities or vibrations interference there will be serious deviation, discrimination meeting degradation, another kind is that set angle threshold value and time threshold judge whether to fall down by Wearable angular transducer human body trunk angle, and the method is difficult to distinguish the normal behaviour action such as bend over, lie low.For example, patent 200620075599.6,200620003000.8 judge whether to fall down by the inclined degree of sensor human body, the very difficult differentiation action such as bend over, lie low, in addition, owing to falling down, event randomness is strong, various informative, therefore, the method False Rate of this threshold decision is higher, and very unstable.
Summary of the invention
Object of the present invention is exactly in order to address the above problem, provide one to fall down detection and location system and method, set up neural network model according to falling over of human body feature, utilize the human body attitude angle that inertial navigation module provides and merge acceleration as neural network input signal, can stablize output by backpropagation, moving window and machine learning algorithm and fall down accurately information, it has accurate detection tested personnel's the situation of falling down, and to its real-time track and localization, system has convenient and practical, accuracy rate is high, the advantage that stability is strong.
To achieve these goals, the present invention adopts following technical scheme:
One is fallen down detection and location system, it comprises falls down detection and location device, the described detection and location device of falling down comprises inertial navigation module, GPS module, master controller and Sim-300 module, described inertial navigation module is made up of three-axis gyroscope, three axle magnetometers and three axis accelerometer, described GPS module, inertial navigation module are connected with master controller respectively, master controller is connected with Sim-300 module, described in fall down detection and location device by GSM/GPRS with binding tested personnel relatives' mobile phone, monitoring station be connected.
Utilize the above-mentioned course of work of falling down detection and location system to be:
Step 1: carry out system initialization to falling down detection and location device, configure the register of each sensor, then set up neural network model according to the feature of falling over of human body, by input teacher signal, network is trained, detailed process is: (1) initialization, suppose not have priori to use, select randomly synaptic weight with consistent a distribution, this distribution is chosen as average and equals 0 be uniformly distributed, the selection of its variance should make the standard deviation of neuronic induction local field be positioned at the linear segment of sigmoid activation function and the transition position of saturated part, (2) collecting sample, AHRS module is worn on to experimenter's waist, experimenter is by making multiple fall down action and normal daily behavior action, record this N group experimental data as training sample (X (N), d (N)), successively each sample (X (i), d (i)) is presented, (3) forward calculation, the training sample of supposing this calculating is (X (n), d (n)), be X (n)={ suma (n), pitch (n), roll (n) }, desired output is d (n), activation function is chosen sigmoid nonlinear function, calculate actual output o (n), error signal e (n)=d (n)-o (n), if meeting the demands, the error signal that each group training sample is obtained through forward calculation (specifies an error range, the error threshold that specifies this neural network is here 0.01%), think that back-propagation algorithm restrains, this group sample training finishes, carry out the training of next group sample, otherwise will carry out backwards calculation, continue to adjust the weights of each layer of neural network, (4) iteration, adjusts after each layer of weights, returns to forward calculation, until error signal meets the convergence criterion of regulation, then, carries out the training of next group sample, until all N group test sample has all been trained,
Step 2: inertial navigation module gathers the gyrostatic signal of 3 axis MEMS, adopt hypercomplex number gesticulate formula, integration is tried to achieve gyroscope attitude angle, gather the signal of 3 axis MEMS accelerometer and three axle magnetometers simultaneously, utilizing the direction cosine conversion between geographic coordinate system and body axis system of gravity field and geomagnetic field to carry out absolute angle resolves, obtain absolute attitude angle, then change in real time filtering parameter according to the change frequency of signal, the attitude angle obtaining for twice is carried out to the data fusion based on extended Kalman filter, wherein, the gyrostatic value of standby accelerometer correction when accelerometer and gyrostatic fusion principle are mainly static, dynamic time, utilize the value of gyrostatic value correction accelerometer, the fusion principle of magnetometer and accelerometer is mainly to utilize accelerometer to carry out the slope compensation of magnetometer, according to the 3-axis acceleration (A of the three axis accelerometer output on movable body
x, A
y, A
z), ask for respectively angle of pitch pitch=tan
-1(Ay ,-Az), pitch angle
merge acceleration
the fusion principle of magnetometer and accelerometer is mainly to utilize accelerometer to carry out the slope compensation of magnetometer, reads the three-axle magnetic field intensity of magnetometer output
ask for the magnetometer output after slope compensation
According to the magnetometer output after the slope compensation calculating.Ask for crab angle
Step 3: master controller receives the data of GPS module and inertial navigation module, attitude angle is inputted to the neural network training with fusion acceleration and fall down judgement, people is in the process of falling down, conventionally the extreme value that merges acceleration sum_a and body obliquity pitch or roll is not to appear in same group of sampled value, in order to improve the discrimination of neural network, farthest distinguish and fall down and daily behavior, adopt the method for moving window to process sampled data, judge according to the Output rusults of neural network, if do not fallen down, master controller uploads to state and positional information by the GPRS communication mode of Sim-300 module the server at monitoring station, if fallen down, master controller is except uploading information monitoring station, also relatives' mobile phone of information notification binding is reminded with note form by the gsm communication mode of Sim-300 module,
Step 4: judge whether this group attitude angle and acceleration information meet the requirement of machine learning, if met, returning to network trains again, upgrade network parameter, if do not met, carry out the processing of next group data, for example, suppose that this group data are once to fall down by a small margin, acceleration and attitude angle change slower, the output valve of neural network is 0.90, this value relatively approaches the threshold value 0.85 of setting, the discrimination of this network that can affect the nerves, adopt the method for machine learning, this input value is trained network again as training sample, network after adjusting is tested again, network output valve is 0.98, this has improved the discrimination of neural network greatly, network is unceasing study in use, accumulate experience, neural network will become more and more intelligence,
Step 5: the server initialization at monitoring station, the information that the device that reception is monitored is uploaded, to depositing database according to No. ID, the form of time of reception, latitude value, longitude, body state after above-mentioned information processing, then database carries out self-inspection, judges whether excess of information, if not, continue reception information, if so, remove expired record, and then reception information;
Step 6: the client at monitoring station is according to authority login account, by Ethernet Query Database, the latitude and longitude information providing according to database, routine call satellite map by the location position of measurand on map, simultaneously according to the status information of measurand, judge whether it falls down, if not, carry out page furbishing, monitor next group information, if so, can start alert program, and then monitor next group information.The obvious advantage of which is that monitoring personnel are not limited to monitoring station, and on the device that can surf the Net at any one, (as mobile phone, computer, online notebook etc.) guard measurand whenever and wherever possible.
Beneficial effect of the present invention:
Native system can detect tested personnel's the situation of falling down exactly, possesses intelligent learning algorithm, be applicable to various types of detections that founder, monitoring personnel can be whenever and wherever possible to tested personnel's real-time follow-up location, and warning device can remind related personnel to give first aid in time, greatly reduces the serious consequence that the elderly or patient cause because of falling down, there is very strong practical value, and system is easy to use, accuracy rate is high, and stability is strong.
Brief description of the drawings
Fig. 1 is the theory diagram of falling down detection system;
Fig. 2 is the workflow diagram of falling down detection system;
Fig. 3 is the operational flowchart at monitoring station;
Fig. 4 is monitoring station client operational flowchart.
Fig. 5 falls down detection algorithm design drawing based on neural network.
Wherein, 1, three-axis gyroscope, 2, three axle magnetometers, 3, three axis accelerometer, 4, inertial navigation module, 5, GPS module, 6, master controller, 7, Sim-300 module, 8, tested personnel's first fall down locating and detecting device, 9, tested personnel's second fall down locating and detecting device, 10, gsm communication, 11, GPRS communication, 12, relatives' mobile phone of binding tested personnel first, 13, relatives' mobile phone of binding tested personnel second, 14, monitoring station, 15, Ethernet.
Embodiment
Below in conjunction with accompanying drawing and embodiment, the invention will be further described.
As Fig. 1, one is fallen down detection and location system, it comprises falls down detection and location device, the described detection and location device of falling down comprises inertial navigation module 4, GPS module 5, master controller 6 and Sim-300 module 7, described inertial navigation module 4 is made up of three-axis gyroscope 1, three axle magnetometers 2 and three axis accelerometer 3, described GPS module 5, inertial navigation module 4 are connected with master controller 6 respectively, master controller 6 is connected with Sim-300 module 7, described in fall down detection and location device and be connected with binding tested personnel's relatives' mobile phone 12, monitoring station 14 by GSM10/GPRS11.
Inertial navigation module 4, by to by three-axis gyroscope 1, three axle magnetometers 2 and three axis accelerometer 3 these three kinds of sensors carry out the data fusion based on EKF, inertial navigation module 4 is exported precise and stable attitude angle and 3-axis acceleration to master controller 6, GPS module 5 is exported accurate latitude and longitude value to master controller 6, master controller 6 completes after neural metwork training, the attitude angle that inertial navigation module 4 is uploaded and 3-axis acceleration adopt the detection algorithm that founders based on neural network and machine learning to carry out data processing, whether can judge exactly this measurand falls down, the GPRS communication mode 11 of master controller 6 by Sim-300 module 7 is by the information of falling down of measurand, together with No. ID of locating information and this device, upload to monitoring station 14, when measurand is while falling down state, the master controller 6 simultaneously GSM short message mode 10 by Sim-300 module 7 sends to the information of falling down on relatives' mobile phone of measurand binding and reminds, the server at monitoring station 14 stores the information of receiving in the database of foundation, the client at monitoring station just can be by the corresponding information in Ethernet 15 Query Databases, can call satellite map and demarcate the accurate location of measurand simultaneously, show the state of falling down of measurand.The detection and location system of falling down of the present invention can be guarded multiple objects simultaneously, and in figure, being tested personnel's first 8 and tested personnel's second 9 taking two guardianships describes as example.Each measurand can upload to information monitoring station 14, they using own No. ID as differentiation, they bind respectively different relatives' mobile phones, and each measurand can be bound multiple mobile phones.
As shown in Figure 2,3, 4, the present invention falls down the workflow of detection and location system: fall down detection and location device 8 and carry out system initialization, configure the register of each sensor.According to the feature modeling neural network of falling over of human body, by input teacher signal, network to be trained, specification error requirement is herein 0.01%.In the time that network output meets error requirements, think that network trained.Inertial navigation module 4 gathers the signal of 3 axis MEMS gyroscope 1, adopts hypercomplex number gesticulate formula, and integration is tried to achieve gyroscope attitude angle.Gather the signal of 3 axis MEMS accelerometer 3 and three axle magnetometers 2 simultaneously, utilize the direction cosine conversion between geographic coordinate system and body axis system of gravity field and geomagnetic field to carry out absolute angle and resolve, obtain absolute attitude angle.Change in real time filtering parameter according to the change frequency of signal, the attitude angle obtaining for twice is carried out to extended Kalman filter, attitude angle and the 3-axis acceleration of final output accurate stable, the neural network that master controller 6 trains attitude angle and acceleration input is fallen down judgement, if do not fallen down, master controller 6 uploads to state and positional information by the GPRS communication mode 11 of Sim-300 module 7 server at monitoring station 14, if fallen down, master controller 6 is except uploading information monitoring station 14, also relatives' mobile phone 12 of information notification binding is reminded with note shape by the gsm communication mode 10 of Sim-300 module 7, finally, judge whether this group attitude angle and acceleration information meet the requirement of machine learning, if met, return to network and again train, and upgrade network parameter, if do not met, carry out the processing of next group data.
After the server initialization at monitoring station 14 completes, start to receive the information that institute's monitoring device is uploaded, to depositing database according to No. ID, the form of time of reception, latitude value, longitude, body state after above-mentioned information processing; Then database carries out self-inspection, judges whether excess of information, if not, continues reception information, if so, removes expired record, and then reception information.
The client at monitoring station 14 is first according to authority login account, by Ethernet Query Database, the latitude and longitude information providing according to database, routine call satellite map by the location position of measurand on map, simultaneously according to the status information of measurand, judge whether it falls down, if not, carry out page furbishing, monitor next group information, if so, can start alert program, and then monitor next group information.
As shown in Figure 4, back-propagation algorithm has been strengthened the reliability of multilayer perceptron supervised training.According to the feature of falling down, first, for input layer, can determine that input layer is 3 neurons: merge acceleration sum_a (variation of acceleration), angle of pitch pitch (trunk forward and backward is fallen down), pitch angle roll (trunk left and right is fallen down).For output layer, owing to only exporting the state (fall down or normally, suppose: o=0 represents that normally, o=1 represents to fall down) of human body here, therefore, output layer arranges 1 neuron.Wherein, XO, YO are respectively the biasings of input layer and hidden layer, get constant+1, and input layer X1, X2, X3 merge acceleration sum_a (n), angle of pitch Pitch (n), pitch angle roll (n),
the synaptic weight that input layer arrives hidden layer,
it is the synaptic weight that hidden layer arrives output layer.Y1, Y2 are hidden nodes, and d (n) is desired output, represent when d (n)=0 normally, represent to fall down when d (n)=1.O (n) is real output value, and e (n) is error signal,
be local gradient, f (g) is activation function.(1) initialization, suppose not have priori to use, we select synaptic weight randomly with consistent a distribution, this distribution is chosen as average and equals 0 be uniformly distributed, and the selection of its variance should make the standard deviation of neuronic induction local field be positioned at the linear segment of sigmoid activation function and the transition position of saturated part.(2) presenting of training sample, AHRS module is correctly worn on to experimenter's waist, experimenter, by making multiple action and some ADL of falling down, records this N group experimental data as training sample (X (N), d (N)).Successively each sample (X (i), d (i)) is presented in order, use for next step calculates.(3) forward calculation, the training sample of supposing this calculating is (X(n), d (n)), be X (n)={ sum_a (n), Pitch (n), roll (n) }, desired output is d (n), activation function is chosen sigmoid nonlinear function, calculate actual output o (n), error signal e (n)=d (n)-o (n), if the error signal that each group training sample is obtained through forward calculation meets the demands, error threshold is 0.01% and thinks that back-propagation algorithm restrains, this group sample training finishes, carry out the training of next group sample, otherwise will carry out backwards calculation, continue to adjust the weights of each layer of neural network, (4) iteration, adjusts after each layer of weights, returns to forward calculation, until error signal meets the convergence criterion of regulation, then, carries out the training of next group sample, until all N group test sample has all been trained.
Claims (2)
1. the method for work of detection and location system is fallen down in a utilization, the described detection and location system of falling down, it comprises falls down detection and location device, the described detection and location device of falling down comprises inertial navigation module, GPS module, master controller and Sim-300 module, described inertial navigation module is by three-axis gyroscope, three axle magnetometers and three axis accelerometer composition, described GPS module, inertial navigation module is connected with master controller respectively, master controller is connected with Sim-300 module, the described detection and location device of falling down is by GSM/GPRS and relatives' mobile phone of binding tested personnel, monitoring station connects, it is characterized in that, concrete steps are:
Step 1: carry out system initialization to falling down detection and location device, configure the register of each sensor, then set up neural network model according to the feature of falling over of human body, by input teacher signal, network is trained, specification error requirement is herein 0.01%; In the time that network output meets error requirements, think that network trained;
Step 2: inertial navigation module gathers the gyrostatic signal of 3 axis MEMS, adopt hypercomplex number gesticulate formula, integration is tried to achieve gyroscope attitude angle, gather the signal of 3 axis MEMS accelerometer and three axle magnetometers simultaneously, utilizing the direction cosine conversion between geographic coordinate system and body axis system of gravity field and geomagnetic field to carry out absolute angle resolves, obtain absolute attitude angle, then change in real time filtering parameter according to the change frequency of signal, the attitude angle obtaining for twice is carried out to the data fusion based on extended Kalman filter, wherein, the gyrostatic value of standby accelerometer correction when accelerometer and gyrostatic fusion principle are mainly static, dynamic time, utilize the value of gyrostatic value correction accelerometer, the fusion principle of magnetometer and accelerometer is mainly to utilize accelerometer to carry out the slope compensation of magnetometer, according to the 3-axis acceleration (A of the three axis accelerometer output on movable body
x, A
y, A
z), ask for respectively angle of pitch pitch=tan
-1(Ay ,-Az), pitch angle
merge acceleration
the fusion principle of magnetometer and accelerometer is mainly to utilize accelerometer to carry out the slope compensation of magnetometer, reads the three-axle magnetic field intensity of magnetometer output
ask for the magnetometer output after slope compensation
According to the magnetometer output after the slope compensation calculating, ask for crab angle
Step 3: master controller receives the data of GPS module and inertial navigation module, attitude angle is inputted to the neural network training with fusion acceleration and fall down judgement, adopt the method for moving window to process sampled data, judge according to the Output rusults of neural network, if do not fallen down, master controller uploads to state and positional information by the GPRS communication mode of Sim-300 module the server at monitoring station, if fallen down, master controller is except uploading information monitoring station, also relatives' mobile phone of information notification binding is reminded with note form by the gsm communication mode of Sim-300 module,
Step 4: judge whether this group attitude angle and acceleration information meet the requirement of machine learning, if met, returning to network trains again, upgrade network parameter, if do not met, carry out the processing of next group data, acceleration and attitude angle change slower, adopt the method for machine learning, this input value is trained network again as training sample, network after adjusting is tested again, network output valve is 0.98, improve the discrimination of neural network, network is unceasing study in use, accumulate experience, neural network will become more and more intelligence,
Step 5: the server initialization at monitoring station, the information that the device that reception is monitored is uploaded, to depositing database according to No. ID, the form of time of reception, latitude value, longitude, body state after above-mentioned information processing, then database carries out self-inspection, judges whether excess of information, if not, continue reception information, if so, remove expired record, and then reception information;
Step 6: the client at monitoring station is according to authority login account, by Ethernet Query Database, the latitude and longitude information providing according to database, routine call satellite map by the location position of measurand on map, simultaneously according to the status information of measurand, judge whether it falls down, if not, carry out page furbishing, monitor next group information, if so, can start alert program, and then monitor next group information.
2. the method for work of detection and location system is fallen down in a kind of utilization as claimed in claim 1, it is characterized in that, the detailed process of described step 1 is: (1) initialization, suppose not have priori to use, select randomly synaptic weight with consistent a distribution, this distribution is chosen as average and equals 0 be uniformly distributed, and the selection of its variance should make the standard deviation of neuronic induction local field be positioned at the linear segment of sigmoid activation function and the transition position of saturated part, (2) collecting sample, AHRS module is worn on to experimenter's waist, experimenter is by making multiple fall down action and normal daily behavior action, record this N group experimental data as training sample (X (N), d (N)), successively each sample (X (i), d (i)) is presented, (3) forward calculation, the training sample of supposing this calculating is (X (n), d (n)), be X (n)={ sum_a (n), pitch (n), roll (n) }, desired output is d (n), activation function is chosen sigmoid nonlinear function, calculate actual output o (n), error signal e (n)=d (n)-o (n), if the error signal that each group training sample is obtained through forward calculation meets the demands, error threshold is 0.01%, think that back-propagation algorithm restrains, this group sample training finishes, carry out the training of next group sample, otherwise will carry out backwards calculation, continue to adjust the weights of each layer of neural network, (4) iteration, adjusts after each layer of weights, returns to forward calculation, until error signal meets the convergence criterion of regulation, then, carries out the training of next group sample, until all N group test sample has all been trained.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201210185040.9A CN102707305B (en) | 2012-06-06 | 2012-06-06 | Tumble detecting and positioning system and method |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201210185040.9A CN102707305B (en) | 2012-06-06 | 2012-06-06 | Tumble detecting and positioning system and method |
Publications (2)
Publication Number | Publication Date |
---|---|
CN102707305A CN102707305A (en) | 2012-10-03 |
CN102707305B true CN102707305B (en) | 2014-07-16 |
Family
ID=46900234
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201210185040.9A Active CN102707305B (en) | 2012-06-06 | 2012-06-06 | Tumble detecting and positioning system and method |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN102707305B (en) |
Families Citing this family (47)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102973274A (en) * | 2012-10-26 | 2013-03-20 | 中国人民解放军空军航空医学研究所 | Posture monitoring system |
CN103142234B (en) * | 2013-01-25 | 2015-04-01 | 中国科学院深圳先进技术研究院 | Method for judging body falldown |
CN103308069B (en) * | 2013-06-04 | 2015-06-24 | 电子科技大学 | Falling-down detection device and method |
CN103377542A (en) * | 2013-07-16 | 2013-10-30 | 中国科学院深圳先进技术研究院 | Human body fall-down preventing early warning method and human body fall-down preventing early warning device |
CN103377541A (en) * | 2013-07-16 | 2013-10-30 | 中国科学院深圳先进技术研究院 | Human body fall-down prevention early warning and intervening method and system |
CN103731557A (en) * | 2014-01-08 | 2014-04-16 | 南方医科大学 | Smartphone-based real-time falling detection system and method |
CN103926608A (en) * | 2014-04-18 | 2014-07-16 | 广州南方卫星导航仪器有限公司 | Incline compensation method of GNSS measuring device |
CN104306003A (en) * | 2014-11-05 | 2015-01-28 | 北京电子科技职业学院 | Foot tracking device |
CN105701965B (en) * | 2014-12-15 | 2020-06-30 | 朱保东 | Wearable safety equipment and method thereof |
CN104825165B (en) * | 2015-03-03 | 2017-11-17 | 北京科技大学 | A kind of intelligent wearable device based on Ubiquitous Network transmission |
CN104814742A (en) * | 2015-05-06 | 2015-08-05 | 中国科学院深圳先进技术研究院 | Health wristwatch and working method thereof |
CN104936063A (en) * | 2015-06-13 | 2015-09-23 | 陈火 | Bluetooth headset |
CN105389762A (en) * | 2015-10-20 | 2016-03-09 | 南京信息工程大学 | Portable evacuation terminal and evacuation realizing method thereof |
CN105193407B (en) * | 2015-10-29 | 2019-04-23 | 上海集成电路研发中心有限公司 | All-weather intelligent monitoring system and Wristwatch type intelligent controller |
CN105528754A (en) * | 2015-12-28 | 2016-04-27 | 湖南师范大学 | Old people information service system based on dual neural network behavior recognition model |
CN105681595A (en) * | 2016-03-30 | 2016-06-15 | 努比亚技术有限公司 | Alarm method and terminal |
CN106056848B (en) * | 2016-06-30 | 2018-11-06 | 深圳先进技术研究院 | A kind of anti-tumble precaution device method for realizing low power consumption |
TWI585435B (en) * | 2016-07-22 | 2017-06-01 | 緯創資通股份有限公司 | Human body positioning method, human body positioning system, and positioning server |
CN106023528A (en) * | 2016-08-01 | 2016-10-12 | 王琰 | Emergency monitoring device for old people |
CN106373335A (en) * | 2016-08-16 | 2017-02-01 | 中国人民解放军海军工程大学 | Inertial technology based and Android phone applied detection and smart alarm method for old people's fall |
CN106228751B (en) * | 2016-08-26 | 2019-06-07 | 陈雷 | Based on Android platform MEMS/ Magnetic Sensor/GPS tumble intelligent alarm system and method |
CN106343975A (en) * | 2016-08-30 | 2017-01-25 | 特斯联(北京)科技有限公司 | Household intelligent medical treatment device |
CN106546261B (en) * | 2016-09-20 | 2019-08-23 | 捷开通讯(深圳)有限公司 | A kind of angle-data compensation method and device based on virtual reality device |
CN106767804B (en) * | 2016-12-28 | 2018-03-27 | 华中科技大学 | The multidimensional data measurement apparatus and method of a kind of moving object |
CN108338790B (en) * | 2017-01-24 | 2020-09-18 | 首都医科大学 | Gait analysis and fall assessment system |
CN107019501B (en) * | 2017-05-05 | 2020-05-05 | 山东师范大学 | Remote tumble detection method and system based on genetic algorithm and probabilistic neural network |
CN107085246A (en) * | 2017-05-11 | 2017-08-22 | 深圳合优科技有限公司 | A kind of human motion recognition method and device based on MEMS |
CN107202594A (en) * | 2017-05-11 | 2017-09-26 | 深圳合优科技有限公司 | A kind of data processing method and device based on MTK platforms |
CN107389063B (en) * | 2017-07-26 | 2020-12-22 | 重庆邮电大学 | High-precision indoor fusion positioning method based on GSM/MEMS fusion |
CN107707657B (en) * | 2017-09-30 | 2021-08-06 | 苏州涟漪信息科技有限公司 | Safety monitoring system based on multiple sensors |
CN107788990A (en) * | 2017-10-26 | 2018-03-13 | 广州大学华软软件学院 | A kind of wearable falling detection device and fall detection system |
CN108151732A (en) * | 2017-12-22 | 2018-06-12 | 浙江西湖高等研究院 | A kind of long-range position and behavior method of estimation |
CN108225271A (en) * | 2018-01-24 | 2018-06-29 | 南京阿凡达机器人科技有限公司 | Method, system, storage medium and the equipment of detection are toppled over by a kind of robot |
CN108338791A (en) * | 2018-02-09 | 2018-07-31 | 张立海 | The detection device and detection method of unstable motion data |
CN108564005B (en) * | 2018-03-26 | 2022-03-15 | 电子科技大学 | Human body falling identification method based on convolutional neural network |
CN108771543B (en) * | 2018-04-16 | 2020-11-03 | 齐鲁工业大学 | Old man falling detection method and system in real environment based on big data |
CN109091151B (en) * | 2018-09-06 | 2021-07-16 | 中国人民解放军战略支援部队信息工程大学 | Pedestrian falling detection method and device based on MIMU |
CN109816933B (en) * | 2019-03-20 | 2020-10-16 | 潍坊医学院 | Intelligent monitoring system and monitoring method for preventing old people from falling down based on composite sensor |
CN110216675B (en) * | 2019-06-20 | 2021-04-06 | 北京猎户星空科技有限公司 | Control method and device of intelligent robot, intelligent robot and computer equipment |
CN110263870B (en) * | 2019-06-26 | 2023-04-18 | 深圳市悦动天下科技有限公司 | Motion state identification method and device, intelligent terminal and storage medium |
CN110889330A (en) * | 2019-10-30 | 2020-03-17 | 西北工业大学 | BP neural network-based old people tumbling detection method and system |
CN111654550A (en) * | 2020-06-16 | 2020-09-11 | 赵继海 | NB-IOT (NB-IOT) -based Internet of things deviation alarm system |
CN112556831B (en) * | 2020-12-01 | 2021-08-27 | 南京艾泰克物联网科技有限公司 | Detection method of intelligent detection alarm device for well lid abnormity |
CN112603296A (en) * | 2020-12-18 | 2021-04-06 | 新绎健康科技有限公司 | Gait analysis method and device based on acceleration sensor |
CN112790760A (en) * | 2021-01-05 | 2021-05-14 | 北京诺亦腾科技有限公司 | Three-dimensional motion attitude capturing method, device, processing equipment and system |
CN113221621B (en) * | 2021-02-04 | 2023-10-31 | 宁波卫生职业技术学院 | Gravity center monitoring and identifying method based on deep learning |
CN117419761B (en) * | 2023-09-27 | 2024-04-19 | 成都天测皓智科技有限公司 | High-precision intelligent sensing refuse landfill situation monitoring method and system |
Citations (8)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
GB2422790A (en) * | 2005-02-07 | 2006-08-09 | Nigel Allister Anstey | Measurement of physical fitness |
CN200941648Y (en) * | 2006-08-07 | 2007-08-29 | 扬州大学 | Cell phone having warning function when dumping short messages |
CN101536053A (en) * | 2006-11-14 | 2009-09-16 | 皇家飞利浦电子股份有限公司 | System for fall prevention and a method for fall prevention using such a system |
CN101579238A (en) * | 2009-06-15 | 2009-11-18 | 吴健康 | Human motion capture three dimensional playback system and method thereof |
CN101650869A (en) * | 2009-09-23 | 2010-02-17 | 中国科学院合肥物质科学研究院 | Human body tumbling automatic detecting and alarming device and information processing method thereof |
CN102023700A (en) * | 2009-09-23 | 2011-04-20 | 吴健康 | Three-dimensional man-machine interactive system |
CN102164532A (en) * | 2008-09-23 | 2011-08-24 | 皇家飞利浦电子股份有限公司 | Power measurement method and apparatus |
CN202600156U (en) * | 2012-06-06 | 2012-12-12 | 山东大学 | Tumbling detection location system |
Family Cites Families (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
WO2007082389A1 (en) * | 2006-01-20 | 2007-07-26 | 6Th Dimension Devices Inc. | Method and system for assessing athletic performance |
-
2012
- 2012-06-06 CN CN201210185040.9A patent/CN102707305B/en active Active
Patent Citations (8)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
GB2422790A (en) * | 2005-02-07 | 2006-08-09 | Nigel Allister Anstey | Measurement of physical fitness |
CN200941648Y (en) * | 2006-08-07 | 2007-08-29 | 扬州大学 | Cell phone having warning function when dumping short messages |
CN101536053A (en) * | 2006-11-14 | 2009-09-16 | 皇家飞利浦电子股份有限公司 | System for fall prevention and a method for fall prevention using such a system |
CN102164532A (en) * | 2008-09-23 | 2011-08-24 | 皇家飞利浦电子股份有限公司 | Power measurement method and apparatus |
CN101579238A (en) * | 2009-06-15 | 2009-11-18 | 吴健康 | Human motion capture three dimensional playback system and method thereof |
CN101650869A (en) * | 2009-09-23 | 2010-02-17 | 中国科学院合肥物质科学研究院 | Human body tumbling automatic detecting and alarming device and information processing method thereof |
CN102023700A (en) * | 2009-09-23 | 2011-04-20 | 吴健康 | Three-dimensional man-machine interactive system |
CN202600156U (en) * | 2012-06-06 | 2012-12-12 | 山东大学 | Tumbling detection location system |
Also Published As
Publication number | Publication date |
---|---|
CN102707305A (en) | 2012-10-03 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN102707305B (en) | Tumble detecting and positioning system and method | |
Shrivastava et al. | Health Monitoring based Cognitive IoT using Fast Machine Learning Technique | |
CN106956271A (en) | Predict the method and robot of affective state | |
CN108245880A (en) | Body-sensing detection method for visualizing and system based on more wearing annulus sensor fusions | |
Kang et al. | Real-time elderly activity monitoring system based on a tri-axial accelerometer | |
CN107533565A (en) | Identified event in the physical data sensed from focus type equipment | |
CN202600156U (en) | Tumbling detection location system | |
Huang et al. | A zigbee-based location-aware fall detection system for improving elderly telecare | |
CN107330240A (en) | A kind of intelligent remote based on both hands ring sensor provides special care to monitoring system and method | |
Xu et al. | Elders’ fall detection based on biomechanical features using depth camera | |
CN110675596A (en) | Fall detection method applied to wearable terminal | |
CN106999748A (en) | The system, apparatus and method relevant with exercise data | |
CN107943276A (en) | Based on the human body behavioral value of big data platform and early warning | |
CN107657277A (en) | A kind of human body unusual checking based on big data and decision method and system | |
Yang et al. | A novel activity recognition system for alternative control strategies of a lower limb rehabilitation robot | |
US20220125359A1 (en) | Systems and methods for automated monitoring of human behavior | |
Tan et al. | Online activity recognition and daily habit modeling for solitary elderly through indoor position-based stigmergy | |
Villar et al. | Autonomous on-wrist acceleration-based fall detection systems: unsolved challenges | |
Yuan et al. | Adaptive recognition of motion posture in sports video based on evolution equation | |
Pham et al. | Human localization and tracking using distributed motion sensors and an inertial measurement unit | |
Chen et al. | Impact of sports wearable testing equipment based on vision sensors on the sports industry | |
Bhargava et al. | Fog-centric localization for ambient assisted living | |
Chen et al. | Improve the Accuracy of Fall Detection Based on Artificial Intelligence Algorithm. | |
Pan et al. | Heterogeneous sensor data fusion for human falling detection | |
Amiroh et al. | Intelligent System for Fall Prediction Based on Accelerometer and Gyroscope of Fatal Injury in Geriatric |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C06 | Publication | ||
PB01 | Publication | ||
C10 | Entry into substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
C14 | Grant of patent or utility model | ||
GR01 | Patent grant |