Detecting Static and Dynamic Objects
AssigneeBayerische Motoren Werke Aktiengesellschaft
InventorGeorg TANZMEISTER
Life of this patent
Estimating the term…
Abstract
For detecting static and dynamic objects, the objects are statistically detected by way of a particle card, and new particles are respectively added in repeating steps. The method is characterized by the fact that static particles are also added, which makes it possible to model static objects in a very precise manner. An environmental model is also generated that has a two-dimensional arrangement of cells, and each cell represents a specific location. To each cell at least two continuous classification values are assigned that describe different attributes of the cells, for example, whether a cell has an object, a static object, a dynamic object, or whether it represents a free space, or whether none of these classes can be assigned.
Claims15
A method for detecting static and dynamic objects, the method comprising the acts of:
- a) distributing new particles in a particle card at a specific time step (t), wherein the particle card is a two-dimensional arrangement of cells and each cell represents a specific location, each particle is depicted by a velocity vector, and the new particles are distributed in the particle card by a random process;
- b) calculating a location of the particles at a later point-in-time (t+Δt) by a predetermined time step (Δt) vis-a-vis the specific point-in-time (t) from step a) and assigning the particles to the cells of the particle card that correspond to the newly calculated location;
- c) measuring a location of real objects via a sensor in an area including the locations of the particle card;
- d) filtering the particles in the individual cells as a function of the objects measured, and in the cells in which no object was measured more particles are deleted than in cells in which an object is located;
- e) repeating steps a) through d) and, in step a), adding new particles to the particles not deleted in step d),
- wherein
- in step a), static particles having zero velocity are also added.
A method for detecting static and dynamic objects, the method comprising the acts of:
- measuring a location of real objects by a sensor;
- generating an environmental model that has a two-dimensional arrangement of cells and each cell represents a specific location, and at least two continuous classification values, describing the probability of a respective cell being assigned to a specific class, are assigned to the cells for respectively one specific time step (t), wherein in particular two or more of the following classes are differentiated as to:
- whether the cell has an object;
- whether the cell has a static object;
- whether the cell has a dynamic object;
- whether the cell represents a free space;
- and
- whether the cell cannot be assigned to any of the aforementioned classes.
A system for detecting static and dynamic objects, comprising a sensor for detecting the location of the objects, and a control device configured to execute a program to:
- a) distribute new particles in a particle card at a specific time step (t), wherein the particle card is a two-dimensional arrangement of cells and each cell represents a specific location, each particle is depicted by a velocity vector, and the new particles are distributed in the particle card by a random process;
- b) calculate a location of the particles at a later point-in-time (t+Δt) by a predetermined time step (Δt) vis-a-vis the specific point-in-time (t) from step a) and assign the particles to the cells of the particle card that correspond to the newly calculated location;
- c) measure a location of real objects via a sensor in an area including the locations of the particle card;
- d) filter the particles in the individual cells as a function of the objects measured, and in the cells in which no object was measured more particles are deleted than in cells in which an object is located;
- e) repeat steps a) through d) and, in step a), add new particles to the particles not deleted in step d),
- wherein
- in step a), static particles having zero velocity are also added.
Dependencies are inferred from the claim text.