Robotics and Gaming Strategy

Academic Papers about Robotics and Gaming Strategy


Motion Planning

Motion planning that is also known as the piano movers problem or the navigation problem is a robotics term that describes the process by which a movement task is broken into discrete motions. Separate motions satisfy movement restrictions and possibly enhance some movement aspect. Consider navigating through a movable robot in a building to some waypoint. The task should be executed while trying to avoid the walls not to fall downstairs. An algorithm for motion planning should take an account of the functions as idea and yield the speed and rotating commands that are sent out to the wheels of the robot. These algorithms address robots with large numbers of joints including industrial manipulators, manipulation of objects, different constraints, and uncertainty.

Motion planning has several applications in robotics that include automation, autonomy, and CAD software robot design. Its other applications are in fields like digital characters, artificial intelligence in video gaming, robotic surgery, and architectural drawings and in the training of organic molecules.

Concepts of Motion planning
A motion planning concept is a problem to yield a uniform incessant motion that tries to connect a start configurations to a figure setting G all this is done while evading collision with hitches that are known. The obstacle and robot geometry is defined in 2D and 3D workspace while the representation of the motion is in a high dimensional path in a configuration.

Configuration Space
A configuration space defines the robot’s posture, and the shape space C encampases all possible formations. Example:
• If it is a zero sized robot forming a 2D plane that is the workplace, C a plane, the shape will be represented by two limits (x,y)
• When the robot forms a 2D shape which rotate and translate, the workstation still is 2D. Though, C is the Euclidian collection SE(2) = R2 x SO(2) here SO(2) forms the superior orthogonal collection in the 2D revolutions. A shape will be presented by (x, y, θ) as the three parameters.
• When the robot has a 3D solid figure that can rotate and translate, the workstation is 3D, and C is the unique Euclidian collection SE(3) = R3 SO(3). A type of formation needs 6 parameters which are Euler angles (α, β, γ) and (x, y, z) for translation.
• Supposing the robot is an immovable base exploiter with a number N revolute joins the C is an N-dimensional oint.

Free Spaces
A set of configuration that tries to avoid crash with hitches is known as a free space the Cfree. Cfree has an accompaniment in C that is known as the forbidden region or the obstacle. It is often hard to calculate the figure of the Cfree is efficient. Forward Kinematics is normally used to fix the location of the geometry of the robot and detection of collision tests if the environments geometry hits with the robots geometry.

Low-dimensional math is resolved with the grid-base procedures which overlay a grip on the geometric algorithms or configuration space that computes the connectivity and shape of Cfree. Careful motion preparation for the high-dimensional structures under the compound constraints is intractable computationally. ThepPotential-field procedures are effectual, but they are an exemption to the harmonic potential fields and are prey to local minima. They do not determine that there is no existence of a path, but the failure probability decreases with the time spent. Currently, sampling-based algorithms are considered for motion planning and have been applied to major problems. One industry that is highly interested in this technology, UK casinos and online gaming for apps, consoles, tablets and smartphones.

Grid-Based Exploration
It approaches a grid happening shape space, and it assumes each shape is recognized by a grid-point. At every grid-point, it moves to neighboring grid points so long as the Cfree is contained in the line

List of Algorithms
• Probabilistic Roadmap Algorithm
• Rapidly exploring Random Tree algorithm
• Sample-Based Algorithms
• Potential Fields
• Geometric Algorithms
• Interval-Based Search

Motion planner is supposed to be comprehensive if infinite time; the planner yields correct reports or a solution that there is none.

Video Games and Robots

Human beings have an important ability robots and computers do not, and that is the ability to discard enormous amounts of data and enormous numbers of choices based on the fact they fail to match a set of intuitive criteria.

The best example of this phenomena is the bread buyer’s dilemma. A seven-year-old child can be told by her mother to go to the store and buy a loaf of bread using only the following two instructions: “Please go to the store and buy a loaf of bread. There is money on the dresser.”

The likelihood of a seven-year-old achieving that task is several orders of magnitude higher than the equivalent probability for a computer or robot. Why?

The first reason is the seven-year-old knows what bread is and what it looks like. She is able to discard everything in the store that is not bread in an instant. Not so for the robot. It must examine everything in the store and compare it to its set of criteria for identifying bread. Although it can do this very quickly, it can’t match the intuitive speed of the child.

The second reason is the human mind’s ability to examine and identify patterns allows it to “fill in the blanks” very quickly. The child can self-author all the other instructions the mother left out and has a high likelihood of getting them right. Again, the robot can’t do this. It can only follow instructions as given and in the order they were given.

Since the mother expects her instructions to be followed in the proper order without specifying that order, she will fail to notice the robot will end up in jail for shoplifting because the location of the money comes after the directive to buy bread. The robot will go to the store first without the money and try to buy bread and fail. Then (assuming it isn’t arrested) it will come back home and get the money. Human beings aren’t used to providing precise instructions because they expect other people to intuit what isn’t immediately clear and to do it correctly.

Researchers are attempting to solve problems like these by using video games ( to train robots to recognize instructions and patterns in those instructions more effectively. Video games give robots and computers a virtualized and consistent probability space where they can experiment with various combinations of instructions, items and results without stealing bread from a supermarket. The goal is to get robots to make the correct associations and then to act on those associations when they are identified later in their experiments.

The ultimate goal is to get robots to learn to apply instructions generally instead of requiring them to be specific to each set of circumstances. It’s an interesting experiment and it should teach us all something about both robots and ourselves.