VR Positioning and Head Tracking in HTC vive, Oculus rift and PlayStation.

Virtual reality(VR ) as the hottest display technology recently, its development prospects and application range is immeasurable. There are three major VR vendors launched their own desktop virtual reality platform Oculus Rift, HTC Vive and Sony PlayStation VR. These three products, whether hardware performance, platform size or resources, have a very high standard. We will summarize and compare the positioning of the three major product technologies in this article.

VR indoor positioning technology can locate VR headset and handle VR devices such as VIVE Tracker in space.  With space-based VR device can not only provide better immersion, the resulting sense of dizziness will be significantly reduced, the entire screen can be a reality, the world is really moving according to our movements. Indoor location technology is very important for VR desktop virtual reality devices.

HTC vive “Lighthouse” is a laser positioning technology, Oculus Rift and Sony PlayStation VR are optical positioning technology. Oculus Rift is the infrared active optical technology and Sony PlayStation VR is the visible light active optical technology.

HTC Vive’s Lighthouse indoor positioning technology.

HTC’s Lighthouse indoor positioning technology is a laser scanning positioning technology, by laser and photosensitive sensors to determine the location of moving objects. Two laser emitters are positioned diagonally to form a rectangular area of adjustable size. The laser beam is emitted by two rows of fixed LED lights inside the transmitter, six times per second. Each laser emitter has two scanning modules, which respectively transmit laser scanning and positioning space to the positioning space in turn in horizontal and vertical directions.

HTC Vive Heads and handles have more than 70 light sensors. By calculating the time to receive the laser to calculate the exact position of the sensor relative to the laser emitter, the position and direction of the visible head can be detected by multiple photosensors. Need to explain here, HTC Vive laser positioning technology used in the positioning process of the photosensitive sensor ID with the data it receives at the same time passed to the computing unit, which means that the computing unit can directly distinguish between different photosensitive sensors, thus A three-dimensional model of the head and handle is finally constructed, based on the position of each light sensor fixed on the head and the handle as well as other information.

Laser positioning technology with low cost, high positioning accuracy, distributed processing and other advantages, and almost no delay, are not afraid to block, even if the handle on the back or crotch can still capture. It can be said that laser positioning technology can achieve indoor positioning with high accuracy and high reaction speed while avoiding the disadvantages of high complexity, high equipment cost, slow computing speed and being easily affected by natural light based on image processing technology. In addition, compared to the other two products, HTC Vive can allow users to carry out activities in a certain space, the user is limited, can adapt to the need to move the game. However, due to the HTC Vive laser launch base station is the use of mechanical control to control the laser scanning positioning space, and mechanical control itself, there is stability and poor durability problems, as a result of HTC Vive stability and durability of slightly worse.

Oculus Rift positioning technology

Oculus Rift uses active optical positioning technology.

Knowing the user of the Oculus Rift may know that the Oculus Rift devices hide some of the infrared light (that is, markers) that emit infrared light and shoot with two infrared cameras in real time. The so-called infrared camera is installed in the infrared filter outside the camera, so the camera can only capture the head and the handle (Oculus touch) on the infrared light to filter out the head and the handle around the visible light signal to improve the image of the letter The noise ratio increases the robustness of the system.

After obtaining the infrared image, the images collected by the two cameras from different angles are transmitted to the computing unit, and the unwanted information is filtered out by the visual algorithm to obtain the position of the infrared light.

Reuse PnP algorithm, that is, the use of four non-coplanar infrared light on the device location information, the four points of the image information can be the final device into the camera coordinate system, fitting the three-dimensional model of the device, and thus Real-time monitoring player’s head, hand movements. What needs to be explained here is that if you want to know the location information of different infrared lamps on the device, you must be able to distinguish between different infrared lamps, the specific solutions are as follows:

It tells the camera’s own ID by blinking the IR lamp. By controlling the camera shutter frequency and flashing frequency of each LED, you can control the size of the image of each infrared lamp on the picture, and then use the 10-frame image of each point in the 10-frame image size changes to determine the LED The corresponding ID number of the light bulb, and then according to the ID number you can know the location of the infrared light on the device information.

In addition, the Oculus Rift is equipped with a nine-axis sensor that uses the nine-axis sensor to calculate the spatial position of the device when infrared optical positioning is occluded or blurred. Due to the obvious zero bias and drift of the nine axes, the infrared optical positioning system can also use the obtained positioning information to calibrate the information obtained by the nine axes when the infrared optical positioning system works normally, so that the infrared optical positioning and the nine axes compensate each other.

Oculus Rift active infrared optics + nine-axis positioning system with high accuracy, strong anti-blocking. Due to the high shooting speed of the cameras used, and because such systems can always obtain the absolute position coordinates of the marker in the current space, there is no cumulative error.
However, due to limited camera viewing, the limited range of products available limits the user’s reach to a great extent, making it impossible to use the Oculus Rift to play virtual reality games that require a wide range of activities such as walking. Therefore, although the Oculus Rift can support simultaneous positioning of multiple targets, the number of targets can not be too much, usually not more than two.

Playstation VR positioning technology

PlayStation VR uses visible active optical positioning technology.

The PlayStation VR device uses a body-sensory camera and a color-illuminated object similar to the previous PS Move to track the position of the human head and controls. Head and handle will put LED lights ball, each handle, head mounted on each one. These LED light balls can emit their own light, and different light ball issued by the light of different colors, so that the camera shooting, the light ball and the background environment, between each ball can be a very good distinction.

PS3 originally used a single camera, by calculating the radius of the ball in the picture to calculate the position of the ball relative to the camera, and ultimately determine the handle and head position. However, the accuracy of a single camera positioning is not high, the robustness is not strong, and sometimes the environment will be identified as a colored object handle, sometimes strong sunlight will not work. Therefore, PS4 uses a somatosensory camera, that is, binocular camera, the use of two images taken by the camera to calculate the three-dimensional coordinates of the ball space. Specific principle: Theoretically speaking, for a point in three-dimensional space, as long as this point can be seen by two cameras at the same time, the two cameras at the same time according to the image and the corresponding parameters, you can determine the point in time In the three-dimensional space location information, as shown below:

After using the somatosensory camera, PS4 positioning accuracy, robustness has been greatly improved.
Determine a good three-dimensional coordinates, that is, x, y, z three degrees of freedom, PS series uses nine to calculate the other three degrees of freedom, and freedom of rotation. Thus, six degrees of freedom are obtained and the spatial position and attitude of the handle are determined.
Through the above description we can know PS can support multiple targets simultaneously located, and in different colors to distinguish. However, due to the poor anti-shelter PS, once the use of multiple occlusions with each other, the positioning immediately affected. And because of the limited range of the binocular camera, PS mobile limited, only available within the scope of camera activity, basically only sit in front of the PC. Although PS4 currently uses a binocular camera, but still using visible light positioning, it is easily affected by the background color. In addition, according to the results of the user experience reflected in the case of faster motion capture camera cannot keep up the issue.
Finally, we can compare the following table to see the advantages and disadvantages of the three product positioning technology.

Positioning accuracyAAAAAAAAAAAA
Stability & DurabilityAAAAAAAAAAA
Anti-light (natural light)AAAAAAAAAAAAA
Multi-target positioningAAAAAAAAAAA
Movable rangeAAAAAAAA

Original source:

Previous: Merry Christmas!

Next: Without the