Patent Grant
11026048
Mobile electronic device users typically like the ability to navigate indoor environments using their mobile electronic devices so that they can find assets and locations easily. Technologies exist that provide indoor location and navigation. However, these technologies are often cost prohibitive to install and utilize.
For example, indoor location and navigation may be achieved by installing a dedicated infrastructure within an indoor environment. However, the costs associated with the hardware and installation of this infrastructure can be high.
As another example, indoor location and navigation may be performed using an existing infrastructure (e.g., trilateration of WiFi access points). However, these techniques generally have a low accuracy and a high latency.
As yet another example, fingerprinting techniques, such as WiFi fingerprinting, may improve the accuracy and latency of an indoor location and navigation system. However, these techniques involve performing walkthroughs of an environment to generate a reference set of fingerprints that correspond to known positions in an indoor environment. Generating this data is time consuming and can be error-prone in certain environments.
This disclosure is not limited to the particular systems, methodologies or protocols described, as these may vary. The terminology used in this description is for the purpose of describing the particular versions or embodiments, and is not intended to limit the scope.
As used in this document, the singular forms “a,” “an,” and “the” include plural references unless the context clearly dictates otherwise. Unless defined otherwise, all technical and scientific terms used in this document have the same meanings as commonly understood by one of ordinary skill in the art. As used in this document, the term “comprising” means “including, but not limited to.”
In an embodiment, a system of determining a location of a mobile electronic device in an indoor environment includes an electronic device and a computer-readable storage medium having one or more programming instructions that, when executed, cause the electronic device to determine an initial position of a mobile electronic device in an indoor environment, determine an initial heading of the mobile electronic device, initialize a set of particles within a threshold distance from the relative location and within a threshold angle from the initial heading, determine a relative location associated with the mobile electronic device based on the initial position, detect a move associated with the mobile electronic device, create a subset of the set of particles based on the move, identify a path that extends from the relative location away from the mobile electronic device at an angle, and determine a first distance between the relative location and a nearest obstacle that is encountered along the path. The system filters the particles in the subset by, for each of the particles in the subset: using a map to determine a second distance between a location of the particle and an obstacle nearest to the particle at the angle, determining a difference between the first distance and the second distance, and assigning a probability value to the particle based on the difference. The system determines whether a deviation of the probability values does not exceed a threshold probability value, in response to determining that the deviation does not exceed the threshold probability value, estimate an actual location of the mobile electronic device, and causes a visual indication of the actual location to be displayed to a user via a display of the mobile electronic device.
The system may determine an initial position of a mobile electronic device in the indoor environment by receiving an initial heading associated with the mobile electronic device obtained by a compass of the mobile electronic device. The system may determine a relative location associated with the mobile electronic device based on the initial position by obtaining the relative location from an augmented reality framework of the mobile electronic device.
The system may create a subset of the set of particles based on the move by, for each of the particles in the set: determining whether the move caused the particle to hit an obstacle as defined by the map, and in response to determining that the move caused the particle to hit an obstacle as defined by the map, not including the particle in the subset.
The system may determine the first distance between the relative location and the nearest obstacle that is encountered along the path by obtaining one or more images of the path that have been captured by a camera of the mobile electronic device, and applying a convolution neural network to one or more of the obtained images to obtain an estimate of the first distance. The convolution neural network may have been trained on the following loss function:
where:
Lprimary is the loss function,
n is a matrix of depth perception estimates,
Yi is an estimated depth perception estimate at position i in n,
Ytrue is an actual distance measurement,
n may have a length of 224.
The system may use the map to determine the second distance between the location of the particle and the obstacle nearest to the particle at the angle by determining a map distance between the location of the particle and the obstacle at the angle on the map, and converting the map distance to the second distance using a scaling factor.
The system may assign a probability value to the particle based on the difference by assigning the probability value to the particle using a Gaussian function.
The system may in response to determining that the deviation exceeds the threshold probability value: create an updated set of particles that include the particles having probability values that exceed a certain value, determine an updated relative location associated with the mobile electronic device, identify an updated path that extends from the updated relative location associated with the mobile device, determine an updated distance between the updated relative location and a nearest obstacle that is encountered along the updated path, and filter the particles in the updated set. The system may filter the particles by, for each of the particles in the updated set: using the map to determine an updated second distance between a location of the particle and one of the plurality of obstacles at the angle, determining a difference between the updated distance and the updated second distance, and assigning an updated probability value to the particle based on the difference.
The system may determine whether a deviation of the updated probability values does not exceed the threshold probability value, in response to determining that the deviation of the updated probability values does not exceed the threshold probability value, estimate an updated actual location of the mobile electronic device, and cause the visual indication of the updated actual location to be displayed to the user via the display of the mobile electronic device.
The system may, in response to determining that the deviation does not exceed the threshold probability value, adjust a heading associated with the mobile electronic device.
The following terms shall have, for purposes of this application, the respective meanings set forth below:
An “electronic device” or a “computing device” refers to a device that includes a processor and memory. Each device may have its own processor and/or memory, or the processor and/or memory may be shared with other devices as in a virtual machine or container arrangement. The memory may contain or receive programming instructions that, when executed by the processor, cause the electronic device to perform one or more operations according to the programming instructions. Examples of electronic devices include personal computers, servers, mainframes, virtual machines, containers, gaming systems, televisions, and mobile electronic devices such as smartphones, personal digital assistants, cameras, tablet computers, laptop computers, media players and the like. In a client-server arrangement, the client device and the server are each electronic devices, in which the server contains instructions and/or data that the client device accesses via one or more communications links in one or more communications networks. In a virtual machine arrangement, a server may be an electronic device, and each virtual machine or container may also be considered to be an electronic device. In the discussion below, a client device, server device, virtual machine or container may be referred to simply as a “device” for brevity.
The terms “memory,” “memory device,” “data store,” “data storage facility” and the like each refer to a non-transitory device on which computer-readable data, programming instructions or both are stored. Except where specifically stated otherwise, the terms “memory,” “memory device,” “data store,” “data storage facility” and the like are intended to include single device embodiments, embodiments in which multiple memory devices together or collectively store a set of data or instructions, as well as individual sectors within such devices.
The term “obstacle” refers to an object or objects that at least partially block, prevent or hinder an individual from traversing a path in an indoor environment. Examples of obstacles include, without limitation, walls, doors, stairways, elevators, windows, cubicles, and/or the like.
The term “particle” refers to a representation of a particular location and/or a heading in an indoor environment.
The terms “processor” and “processing device” refer to a hardware component of an electronic device that is configured to execute programming instructions. Except where specifically stated otherwise, the singular term “processor” or “processing device” is intended to include both single-processing device embodiments and embodiments in which multiple processing devices together or collectively perform a process.
In an embodiment, a remote electronic device 102a-N may be located remotely from a mobile electronic device 100. A server is an example of a remote electronic device 102a-N according to an embodiment. A remote electronic device 102a-N may have or be in communication with one or more data stores 104.
A mobile electronic device 100 may be in communication with one or more remote electronic devices via one or more communication networks 106. A communication network 106 may be a local area network (LAN), a wide area network (WAN), a mobile or cellular communication network, an extranet, an intranet, the Internet and/or the like.
A mobile electronic device 100 may include one or more sensors that provide compass functionality. For instance, a mobile electronic device 100 may include a magnetometer 108. A magnetometer 108 may measure the strength and direction of magnetic fields, which may permit a mobile electronic device 100 to determine its orientation.
A mobile electronic device may include one or more cameras 112. As discussed below, a camera may be an RGB (Red, Green, Blue) camera, an RGB-D camera, and/or the like.
In various embodiments, a mobile electronic device 100 may support an augmented reality (AR) framework 114. An AR framework 114 refers to one or more programming instructions that when executed, cause a mobile electronic device to perform one or more actions related to integrating digital content into a real-world environment. In this document, the term “augmented reality” or “AR” when used with reference to an electronic device or method of using an electronic device, refers to the presentation of content so that the user of the device is able to see at least part of the real-world environment with virtual content overlaid on top of the real-world environment. A mobile electronic device 100 that supports an AR framework 114 may cause virtual content to be overlaid on top of a real-world environment as depicted through a camera application. For example, a camera 112 of a mobile electronic device 100 may capture one or more images of a real-world environment, and an AR framework 114 may cause virtual content to be overlaid on top of these images.
As illustrated in
In various embodiments, a wireless access point 110 may log the time and the strength of one or more communications from a mobile electronic device 100. The wireless access point 110 may send at least part of the logged information to an electronic device such as, for example, a remote electronic device 102a-N. The remote electronic device 102a-N may use the received information to estimate a location of a mobile electronic device 100. For example, a remote electronic device 102a-N may use the received information to determine a position of a mobile electronic device 100 relative to a fixed point in the environment. A remote electronic device may store or have access to a map of a relevant environment, and may use the map to determine a position of a mobile electronic device relative to a reference point. This position may be measured as a certain distance from a reference point, or as one or more position coordinates, such as longitude and latitude.
In various embodiments, an indoor location tracking system, such as the one described with respect to
An indoor location tracking system may use information from an AR framework of a mobile electronic device being tracked to establish a relative distance and heading. A depth estimation technology may provide information about distances from the mobile electronic device to one or more obstacles. An indoor location tracking system may utilize a particle filter to fuse together data to provide an indoor location and heading estimate for the mobile electronic device.
In various embodiments, an indoor location tracking system may determine 202 a start heading associated with the mobile electronic device. For example, one or more sensors of the mobile electronic device (e.g., a magnetometer) may obtain a start heading associated with the mobile electronic device. The obtained start heading may be within twenty degrees of the true heading of the mobile electronic device in various embodiments.
An indoor location tracking system may initialize 204 one or more particles around the start location and start heading for the mobile electronic device. A particle refers to a representation of a particular location and/or a heading in the indoor environment.
An indoor location tracking system may initialize 204 particles within a threshold distance from the start location. For instance, the system may initialize 204 particles +/−50 feet from the start location (e.g., (start x, start y) position). Other threshold distances may be used within the scope of this disclosure. An indoor location tracking system may initialize 204 particles within a threshold angle relative to the start heading. For example, the system may initialize 204 one or more particles within +/−20 degrees from the start heading.
In various embodiments, the system may generate 206 a subset of the initialized particles. The subset may be generated 206 based on a position of the initialized particles. For instance, the system may determine whether any of the initialized particles have a position that corresponds to a position of one or more obstacles as defined by a map of an indoor environment, as discussed in more detail below. The system may generate 206 a subset of particles that excludes these particles.
An indoor location tracking system may determine 208 a relative location and a relative yaw value associated with the mobile electronic device. In various embodiments, an indoor location tracking system may obtain 208 a relative location and/or a relative yaw value from an AR framework associated with the mobile electronic device. A relative location refers to a current location of a mobile electronic device relative to its start location. A relative location of a mobile electronic device may be represented as coordinates such as, for example, (x, y). A relative yaw value refers to a yaw value relative to a start yaw value.
For example, an AR framework may access a camera of a mobile electronic device to obtain one or more images of an indoor environment. The AR framework may perform one or more image processing techniques on the image(s) to determine a relative location and/or a relative yaw value associated with the electronic device. Alternatively, an AR framework may determine a relative location and/or relative yaw associated with an electronic device based on motion information captured by one or more sensors of the mobile electronic device such as, for example, a gyroscope, an accelerometer and/or the like.
Referring back to
Referring back to
An indoor location tracking system may identify 216 one or more target angles, each referred to in this document as a theta. Each target angle may be within a certain range of the relative yaw value. For example, a theta may be within 20 degrees from the relative yaw value. Additional and/or alternate ranges may be used within the scope of this disclosure.
For each of the identified target angles, the indoor tracking system may determine 218 a distance between a relative location of the mobile device and an obstacle nearest to the relative location at the target angle (referred to in this disclosure as a mobile device distance). In various embodiments, an indoor tracking system may identify a path that extends away from the relative location of the mobile electronic device at the target angle. The system may determine a distance between the relative location and the first (or nearest) obstacle that is encountered along the path.
As an example, if a relative location of a mobile electronic device is represented by (A, B) and the target angle is 15 degrees, the indoor tracking system may determine a distance between (A, B) and obstacle at 15 degrees.
In various embodiments, the system may determine 218 mobile device distance relative to an obstacle. A camera associated with a mobile electronic device may capture one or more images of its surrounding environment. In various embodiments, the camera may be a monocular RGB (Red, Green, Blue) camera. The camera may be a RGB-D camera, which may include one or more depth-sensing sensors. The depth sensor(s) may work in conjunction with a RGB camera to generate depth information related to the distance to the sensors on a pixel-by-pixel basis. A camera may be integrated into the mobile electronic device such as, for example, a rear-facing and/or a front-facing camera. In other embodiments, a camera may be one that is attached to or otherwise in communication with a mobile electronic device.
The system may obtain one or more of the captured images from the camera, and may apply a machine learning model such as, for example, a convolutional neural network (CNN), to one or more of the obtained images 700 to determine a depth estimate between the mobile electronic device and an obstacle. A CNN may be pre-trained using a set of color images. A CNN may be used to extract image features separate from depth and color modalities, and subsequently combine these features using a fuser technique.
As illustrated by
For each convolutional layer 702a-N, a set of parameters may be initialized in the form of an array or matrix (referred to in this disclosure as a kernel). The kernel may be applied across a width and height of an input image to convolve the parameters with brightness intensities for the pixels in the input image subject to a threshold for each pixel to generate a feature map having a dimensionality. Each convolution may represent a neuron that looks at only a small region of an input image based on the applied kernel. The number of neurons outputted from a convolution layer may depend on the depth of the applied kernel. A subsequent convolutional layer may take as input the output of a previous convolutional layer and filters it with its own kernel.
In various embodiments, convolutional layers 702a-N may be combined with one or more global average pooling (GAP) layers 704a-N. A GAP layer may calculate the average output of each feature map in the previous layer. As such, a GAP layer 704a-N may serve to significantly reduce the data being analyzed and reduce the spatial dimensions of a feature map.
The output of the GAP layers 704a-N may be provided to a fully-connected layer 706. This output may be represented as a real-valued array having the activations of only a predetermined number of neurons. For instance, the output may be represented as an array of depth estimates 708 for one or more obstacles of an input image.
As an example, applying a CNN to images denoting one or more obstacles may generate a one-dimensional array of depth perception estimates. The array may include one or more angle-distance pairs. An angle value of an angle-distance pair may represent an angle of an obstacle relative to a camera, for example a camera of a mobile electronic device that captured one or more of the images. A distance value of an angle-distance pair may represent an estimated distance between the camera and an obstacle at the corresponding angle. The array may have a length of 224. However, it is understood that alternate lengths may be used within the scope of this disclosure.
In various embodiments, a CNN may be trained on a loss function. An example of such a loss function may be represented by the following:
This loss function penalizes the bigger errors more than the smaller ones, and helps to stabilize the root mean square error while training. It is understood that other loss functions may be used within the scope of this disclosure.
In various embodiments, a CNN may be fine-tuned based on the following function:
It is understood that other functions may be used to fine tune a CNN.
In various embodiments, the system may utilize one or more CNNs to determine a confidence metric associated with one or more of the depth perception estimates described above. In an embodiment, the CNN may be the same CNN as discussed above with respect to
A confidence metric refers to an indication of the accuracy of a depth perception estimate. For instance, a confidence metric may be a value or a range of values that are indicative of a confidence that an associated depth perception estimate is accurate.
For each convolutional layer 802a-N, a set of parameters may be initialized in the form of an array or matrix (referred to in this disclosure as a kernel). The kernel may be applied across a width and height of an input image to convolve the parameters with brightness intensities for the pixels in the input image subject to a threshold for each pixel to generate a feature map having a dimensionality. Each convolution may represent a neuron that looks at only a small region of an input image based on the applied kernel. The number of neurons outputted from a convolution layer may depend on the depth of the applied kernel. A subsequent convolutional layer may take as input the output of a previous convolutional layer and filters it with its own kernel.
In various embodiments, convolutional layers 802a-N may be combined with one or more global max pooling (GMP) layers 804a-N. A GMP layer may calculate the maximum or largest output of each feature map in the previous layer.
The output of the GMP layers 804a-N may be provided to a confidence layer 806. This output may be represented as a confidence metric. For instance, an example of a confidence metric may be a value between ‘0’ and ‘1’, where values closer to ‘0’ indicate a low confidence and values closer to ‘1’ indicate a high confidence. In various embodiments, applying a CNN may generate a one-dimensional array of confidence values that may correspond to one or more depth perception estimates. As such, a confidence value may indicate an estimated measure of how accurate a depth perception estimate is.
In various embodiments, the system may not update a machine learning model to incorporate a depth perception estimate into if the confidence metric associated with the depth perception estimate is lower than a threshold value, is outside of a range of threshold values, and/or the like. For instance, if confidence metrics have values between ‘0’ and ‘1’, the system may not update a machine learning model to incorporate a depth perception estimate if the confidence metric associated with the depth perception estimate is lower than 0.80. Additional and/or alternate confidence value ranges and/or threshold values may be used within the scope of this disclosure.
For one or more of the particles in the subset, the indoor tracking system may determine 220 a distance between the particle's location and a nearest obstacle at one or more of the identified target angles (referred to in this disclosure as a particle distance).
The indoor tracking system may determine 220 a distance between a particle's location and an obstacle depicted on the map at one or more of the identified target angles. The system may identify a path that extends away from the particle's location at a target angle. The system may determine a distance between the particle's location and the first (or nearest) obstacle that is encountered along the path.
For instance, referring to the example above, the indoor tracking system may determine a distance between each particle's location and a nearest obstacle at one or more of the identified target angles illustrated in Table 1.
Examples of such distances for three example particles are illustrated below in Table 2.
The indoor tracking system may determine 220 a distance between a particle's location and an obstacle depicted on the map at one or more of the identified target angles by measuring a distance between the particle's location and a first obstacle that is encountered at the particular target angle on the map. For example,
The indoor tracking system may convert 222 the determined distance into an actual distance. The indoor tracking system may convert 222 the determined distance into an actual distance by applying a scaling factor to the determined distance. The scaling factor may be stored in a data store of the indoor tracking system, or a data store associated with the indoor tracking system.
For example, a quarter of an inch on a map may translate to a distance of one foot in the real environment. As such, if a distance between a particle's location and an obstacle is one inch on the map, the actual distance may be determined to be four feet. Additional and/or alternate scaling factors may be used within the scope of this disclosure.
In various embodiments, the indoor tracking system may determine 224 a difference between the mobile device distance at a theta and a particle distance for one or more of the particles at the theta. For instance, referring to the above example, Table 3 illustrates the mobile device distance, particle distance, and difference between the two for each theta.
The indoor tracking system may convert 226 one or more of the distance values to a probability value. In various embodiments, the indoor tracking system may convert 226 one or more of the distance values to a probability value using any suitable probability distribution such as, for example, a Gaussian function.
The indoor tracking system may resample 228 particles based on their probability values. For instance, the system may select particles having a probability value that is within a certain value range or that exceeds a threshold value. The system may discard the other particles. As such, particles whose distance error is relatively small are more likely to be retained in the resampling.
In various embodiments, the system may determine 230 a deviation associated with the probabilities of the particles in the resampling. A deviation may be a measure of the dispersion of the probabilities relative to one or more certain values. For instance, a deviation may be a standard deviation of the probabilities of the particles in the resampling. Additional and/or alternate deviations may be used within the scope of this disclosure.
If the deviation is not less than a threshold value, the system may repeat steps 208-230 using the resampling. In various embodiments, the system may repeat steps 208-230 until the deviation of the probabilities associated with the particles in the resampling converge. The deviation of the probabilities associated with particles in a resampling may converge when it becomes less than a threshold value.
In response to the deviation of the probabilities converging, the system may optionally adjust 232 the heading of the mobile electronic device. If the error associated with the start heading determination is too high, this may result in a failed path associated with the mobile electronic device. A failed path may be a path or trajectory that is not feasible for an individual or a mobile electronic device to follow. For instance, a failed path may be one that passes through one or more obstacles.
To compensate for potentially high error associated with the start heading, the system may adjust 232 the heading. The system may adjust 232 the heading by traversing data sets associated with a failed path in a forward and/or a backward direction for example, by utilizing a forward-backward propagation strategy.
A position of the mobile electronic device may change 1102. For example, a user of the mobile electronic device may move or otherwise change position. In various embodiments, the indoor location tracking system may create 1104 a subset of particles. The system may determine whether the move has caused one or more of the particles in the current particle to hit an obstacle as indicated by the map. If the system determines that the move has caused a particle to hit an obstacle, the system may not include the particle in the subset. As such, the subset of particles that is created 1104 by the system only includes those particles that the move has not caused to hit an obstacle.
The system may then resample 1106 the subset. In various embodiments, the system may randomly sample particles from the subset as part of the resampling. The system may repeat steps 1100-1106 forwards and/or backwards along the failed path in order to adjust the heading of the mobile electronic device.
In various embodiments, the system may estimate 234 an actual location and/or heading of the mobile electronic device based on the resampling. In various embodiments, a system may estimate 234 an actual location and/or heading of the mobile electronic device by determining a metric associated with tat least a portion of the particles in the resampling. For example, in an embodiment, the system may estimate 234 an actual location of the mobile electronic device by determining a mean location value or a median location value of the locations of the particles in the resampling. Similarly, the system may estimate 234 an actual heading of a mobile electronic device by determining a mean heading value or a median heading value of the headings of the particles in the resampling.
In various embodiments, the system may adjust an estimated location of the mobile electronic device. The system may adjust an estimated location of the mobile electronic device if the estimated location corresponds to an obstacle on the map. For instance, the system may determine an estimated location, which corresponds to a wall on the map. The system may adjust the estimated location so that the location does not conflict with an obstacle. For instance, the system may determine the nearest location to the estimated location that does not conflict with an obstacle, and may adjust the estimated location to this position.
The system may cause 236 a visual depiction of at least a portion of the map to be displayed on a graphical user interface of the mobile electronic device. The visual depiction may include a visual indication of the estimated actual location on the map. The visual indication may include, for example, a colored dot, a symbol, an image, or other indicator.
As illustrated by
A memory device 1220 is a hardware element or segment of a hardware element on which programming instructions, data, or both may be stored. Read only memory (ROM) and random access memory (RAM) constitute examples of memory devices, along with cloud storage services.
An optional display interface 1230 may permit information to be displayed on the display 1235 in audio, visual, graphic or alphanumeric format. Communication with external devices, such as a printing device, may occur using various communication devices 1240, such as a communication port or antenna. A communication device 1240 may be communicatively connected to a communication network, such as the Internet or an intranet.
The hardware may also include a user input interface 1245 which allows for receipt of data from input devices such as a keyboard or keypad 1250, or other input device 1255 such as a mouse, a touch pad, a touch screen, a remote control, a pointing device, a video input device and/or a microphone. Data also may be received from an image capturing device 1210 such as a digital camera or video camera. A positional sensor 1215 and/or motion sensor 1265 may be included to detect position and movement of the device. Examples of motion sensors 1265 include gyroscopes or accelerometers. An example of a positional sensor 1215 is a global positioning system (GPS) sensor device that receives positional data from an external GPS network.
The features and functions described above, as well as alternatives, may be combined into many other different systems or applications. Various alternatives, modifications, variations or improvements may be made by those skilled in the art, each of which is also intended to be encompassed by the disclosed embodiments.
| Number | Name | Date | Kind |
|---|---|---|---|
| 7958562 | Gaucas | Jun 2011 | B2 |
| 8548738 | Kadous et al. | Oct 2013 | B1 |
| 8583400 | Thrun et al. | Nov 2013 | B2 |
| 8934112 | Gross et al. | Jan 2015 | B1 |
| 8982384 | Evanitsky | Mar 2015 | B2 |
| 8986867 | Baek | Mar 2015 | B2 |
| 9201619 | Gross et al. | Dec 2015 | B2 |
| 9367271 | Zehler et al. | Jun 2016 | B2 |
| 9549089 | Tredoux et al. | Jan 2017 | B1 |
| 9894604 | Tran et al. | Feb 2018 | B1 |
| 10440221 | Ebner | Oct 2019 | B2 |
| 20040075861 | Shima et al. | Apr 2004 | A1 |
| 20040264427 | Jaakkola et al. | Dec 2004 | A1 |
| 20060221901 | Yaqub et al. | Oct 2006 | A1 |
| 20070019670 | Falardeau | Jan 2007 | A1 |
| 20080080458 | Cole | Apr 2008 | A1 |
| 20090028082 | Wynn et al. | Jan 2009 | A1 |
| 20100040029 | Doppler et al. | Feb 2010 | A1 |
| 20110063663 | Kim et al. | Mar 2011 | A1 |
| 20110306364 | Gossain et al. | Dec 2011 | A1 |
| 20120075664 | Nichols et al. | Mar 2012 | A1 |
| 20120179737 | Baranov et al. | Jul 2012 | A1 |
| 20130286942 | Bonar et al. | Oct 2013 | A1 |
| 20140022587 | Coccia et al. | Jan 2014 | A1 |
| 20140256286 | Rangarajan | Sep 2014 | A1 |
| 20150189023 | Kubota et al. | Jul 2015 | A1 |
| 20150363141 | Fernandes et al. | Dec 2015 | A1 |
| 20170123737 | Januszewski et al. | May 2017 | A1 |
| 20170171718 | Jeong | Jun 2017 | A1 |
| 20170289813 | Pashkov et al. | Oct 2017 | A1 |
| 20170347388 | Cai et al. | Nov 2017 | A1 |
| 20180083795 | Zehler et al. | Mar 2018 | A1 |
| 20180172451 | Wang | Jun 2018 | A1 |
| 20180237137 | Tovey et al. | Aug 2018 | A1 |
| 20180316821 | Ebner | Nov 2018 | A1 |
| 20190373407 | Bhatti | Dec 2019 | A1 |
| Number | Date | Country |
|---|---|---|
| 103925923 | Jun 2017 | CN |
| 2009187103 | Aug 2009 | JP |
| Entry |
|---|
| Information about Related Patent and Patent Applications, see section 6 of the accompanying Information Disclosure Statement Letter, which concerns Related Patents and Patent Applications. |
| Saxena, A. et al., “Learning Depth from Single Monocular Images”. |
| Godard, C. et al., “Digging Into Self-Supervised Monocular Depth Estimation”. |
| Bhoi, A., “Monocular depth Estimation: A Survey”, [cs.Cv] Jan. 27, 2019. |
| Wu, Y. et al., “HTrack: An Efficient Heading-Aided Map Matching for Indoor Localization and Tracking”, IEEE Sensors Journal, vol. 19, No. 8, Apr. 15, 2019. |
| Seco, F. et al., “Smartphone-Based Cooperative Indoor Localization with RFID Technology”, Sensors 2018, 18, 266. |
| Murata M. et al., “Smartphone-based Indoor Localization for Blind Navigation across Building Complexes”, 2018 IEEE International Conference on Pervasive Computing and Communications (PerCom). |
