I'm developing a GPS system and to do that I'd like to use the A* algorithm. I have a graph where the vertex are the source / target and the edges are the streets. For that I have a database with following info:
id;"Street Name";source;target;GeoCoordinateX1;GeoCoordinateY1;GeoCoordinateX2;GeoCoordinateY2
Each of these lines represent an edge. Using the coordinates the aim is using one path finding algorithm get the shortest and fastest path. I already developed djikstra algorithm but now I'm trying to find a really good heuristic.
I'd like to know if there's an heuristic that is more accurate or efficient. I read that I could use Manhattan, Euclidian or Diagonal distance. I think Euclidian would be a good choice but then the cost of the g funciton wouldnt be the same of the heuristic funcion h. I'd get the shortest path but it'd take longer. Is there any way to get the good and the profit?
Best regards
For most coordinate based pathfinding, the cost function used in A* is 'distance'. Each vertex in your graph is a gps coordinate along a street, so the distance between each vertex is calculated like this using the Haversine formula:
function getDistanceFromLatLonInKm(lat1,lon1,lat2,lon2) {
var R = 6371; // Radius of the earth in km
var dLat = deg2rad(lat2-lat1); // deg2rad below
var dLon = deg2rad(lon2-lon1);
var a = Math.sin(dLat/2) * Math.sin(dLat/2) +
Math.cos(deg2rad(lat1)) * Math.cos(deg2rad(lat2)) *
Math.sin(dLon/2) * Math.sin(dLon/2)
;
var c = 2 * Math.atan2(Math.sqrt(a), Math.sqrt(1-a));
var d = R * c; // Distance in km
return d;
}
function deg2rad(deg) {
return deg * (Math.PI/180)
}
In order to find the fastest path, the cost function will be 'estimated travel time'. In order to calculate this, you need the 'speed limit' for each edge between vertices (street), as well as the distance calculated above.
The cost function is very simple: cost in hours = distance / speed = km / (km/h)
Google maps has an API to find the shortest/fastest paths for many different modes of transportation, and it will save you a lot of effort. It also takes into account traffic, bus schedules and other factors that would be difficult to consider in your project.
If this is a hobby project, continue and you'll learn a lot about pathfinding. Feel free to learn from my favourite A* guide on Amit Patel's website. If you'd like to learn about all the cool options you have for pathfinding, I have an in-depth guide that I wrote for this answer. People in academia commonly write tutorials on the essential algorithms and google searches can find research papers about variants on these algorithms.
Related
I have a constant acceleration and I need to calculate the instantaneous speed and the travelled distance.
So we know that the speed is the integral of acceleration and the travelled distance is the integral of the speed.
Using systemC AMS with the TDF modeling formalism, there isn't the integral operator, but there is the Laplace transform operator instead.
The schema for the speed and distance calculation that I would realize is the follow:
NB:
// Declaration of private variables of SCA_TDF_MODULE
sca_tdf::sca_ltf_nd lft;
sca_util::sca_vector<double> num, den;
// initialisation of variables
num(0) = 1.0;
den(0) = 1.0;
So I thought that I could calculate the speed using the sca_tdf::sca_ltf_nd in this way:
if (throttle_is_active) {
speed.write(lft(num, den, 2.78)); // speed is an output port
// distance ....
}
But I'm not sure that is right, and I don't know how to save this value to calculate the travelled distance using the ltf again.
The definition of tdf module can be found here.
Anyone can help me please?
see also SystemC AMS Users Guide provided with the proof-of-concept library
The coefficients in SystemC-AMS are defined as follows:
H(s) = (num(0) + num(1)*s + num(2)*s**2 ...) / (den(0) + den(1)*s + den(2)*s**2 ...)
-> the coefficients for an integrator are:
num(0)=1.0;
den(0)=0.0;
den(1)=1.0;
I need to calculate the distance between GPS co-ordinates to calculate distance being traveled. I've tried both the Haversine and Vincenty algorithms, which work fine on my desktop PC, but when I use the same code for MSP430 chip using CCS IDE, the IDE is throwing error saying that "program will not fit into available memory".
Is there any other alternative method or code to find the distance between two GPS co-ordinates? which will fit in available memory of MSP430 Microcontroller ?
It's not surprising that you're running out of memory, because the microcontroller you are using, the Texas Instruments MSP430F2274, has only 32kB of flash, and 1kB of RAM.
There are several approaches to solving your problem, each with different tradeoffs. Here are three:
Use another microcontroller that has more memory (there are many in the MSP430 family).
Optimize your code to fit in the available space.
Use a simpler formula than the Vincenty or Haversine.
I'll address the two latter approaches below.
Optimize Your Code
Depending on the accuracy requirements of your application, optimizing your existing code might be a better approach than using a simpler formula than Vincenty or Haversine.
A Simple Way to Optimize
Perhaps simply setting the compiler to optimize for size will solve your problem. In the MSP430 toolset, use the --opt_for_speed=0 switch. According to the MSP430 Optimizing C/C++ Compiler User's Guide (v15.9.0.STS) (page 62), this switch:
enables optimizations geared towards improving the code size with a
high risk of worsening or impacting performance.
So you might very easily get things to work by using this switch, at the cost of trading away speed for memory space.
A More Involved Way to Optimize
Assuming you are using the floating point math library provided with your compiler, you might be able to still use Vincenty or Haversine if you replace the math library with a more space-efficient version. The CORDIC fixed-point algorithms provide iterative approaches to calculating the trigonometric functions, that trade away speed for space efficiency. If you roll your own math library, you might achieve a good balance between space, speed, and accuracy. A 16-bit version of the CORDIC approach for sine() and cosine() for the MSP430 is here; you will need to determine whether it provides the degree of accuracy and precision you need.
Use a Different Formula
In general, the various algorithms that calculate distance between two points on the earth represent a trade-off between accuracy and complexity. The Vincenty algorithm you cited is much more accurate than the Haversine, as it more correctly represents the earth as an oblate spheroid instead of as a sphere of radius R; hence the math is more complex.
For reference, the Haversine method (which assumes the earth is a perfect sphere) is shown here:
dlon = lon2 - lon1
dlat = lat2 - lat1
a = sin^2(dlat/2) + cos(lat1) * cos(lat2) * sin^2(dlon/2)
c = 2 * arcsin(min(1,sqrt(a)))
d = R * c
The intermediate result c is the distance in radians. The distance d is in the same units as R (radius of the earth).
As you can see, the Haversine employs an arcsin() in the calculation.
You can simplify the math further by employing the Polar Coordinate Flat-Earth method:
a = pi/2 - lat1
b = pi/2 - lat2
c = sqrt(a^2 + b^2 - 2 * a * b * cos(lon2 - lon1)
d = R * c
Notice that there is no arcsin() in this calculation, but there is a sqrt().
A discussion of the accuracy tradeoffs between the Haversine and the Polar Coordinate Flat-Earth methods is here, question Q5.1.
See also
How do I calculate distance between GPS co-ordinates on a processor with poor floating point support?
http://www.faqs.org/faqs/geography/infosystems-faq/ Question Q5.1
Geographical Distance (Wikipedia)
After some time searching, I have revised my question.
I have found numerous examples of ball to ball collisions, but the only ones that seem to work use Vector2d or Vector2D.
This is a problem, because I am only allowed to use the regular java library, so my main question is: How do I convert the examples (which I will post below) to use what I can use?
I have several variables, both balls have the same mass, the velocities are broken into different variables, x and y. Also I have access to their x and y pos.
This is the ONLY problem left in my application.
I am at a total loss on how to convert the below example.
// get the mtd
Vector2d delta = (position.subtract(ball.position));
float d = delta.getLength();
// minimum translation distance to push balls apart after intersecting
Vector2d mtd = delta.multiply(((getRadius() + ball.getRadius())-d)/d);
// resolve intersection --
// inverse mass quantities
float im1 = 1 / getMass();
float im2 = 1 / ball.getMass();
// push-pull them apart based off their mass
position = position.add(mtd.multiply(im1 / (im1 + im2)));
ball.position = ball.position.subtract(mtd.multiply(im2 / (im1 + im2)));
// impact speed
Vector2d v = (this.velocity.subtract(ball.velocity));
float vn = v.dot(mtd.normalize());
// sphere intersecting but moving away from each other already
if (vn > 0.0f) return;
// collision impulse
float i = (-(1.0f + Constants.restitution) * vn) / (im1 + im2);
Vector2d impulse = mtd.multiply(i);
// change in momentum
this.velocity = this.velocity.add(impulse.multiply(im1));
ball.velocity = ball.velocity.subtract(impulse.multiply(im2));
Here is the URL for the question:
http://stackoverflow.com/questions/345838/ball-to-ball-collision-detection-and-handling
And I have taken a look at his source code.
Thank you for taking the time to read this issue.
SUCCESS!
I have found how to use Vector2d, and it works PERFECTLY!
Will edit later with answer!
I'm implementing my own 3d engine in c# based on a really basic 3d open-source engine in JavaScript called a3. I don't know If I have 100% understand you but It sounds like you can only find examples with Vector2d but you are not allowed to use that class?
I that is the case, as you can imagine javascript does not have native Vector2d types so someone had to implement. Don't be afraid of giving it a try, is just a few high school maths functions, you should be able to implement your own Vector2d class in just a few minutes
The following link contain implementations if vector2d, vector3d, vector4d, matrix3, and matrix4 in javascript: https://github.com/paullewis/a3/tree/master/src/js/core/math hope it helps :)
I have a database of user submitted latitude/longitude points and am trying to group 'close' points together. 'Close' is relative, but for now it seems to ~500 feet.
At first it seemed I could just group by rows that have the same latitude/longitude for the first 3 decimal places (roughly a 300x300 box, understanding that it changes as you move away from the equator).
However, that method seems to be quite lacking. 'Closeness' can't be significantly different than the distance each decimal place represents. It doesn't take into account that two locations may have different digits in the 3rd (or any) decimal place, but still be within the distance that place represents (33.1239 and 33.1240).
I've also mulled over the situation where Point A, and Point C are both 'close' to Point B (but not each other) - should they be grouped together? If so, what happens when Point D is 'close' to point C (and no other points) - should it be grouped as well. Certainly I have to determine the desired behavior, but how would either be implemented?
Can anyone point me in the right direction as to how this can be done and what different methods/approaches can be used?
I feel a bit like I'm missing something obvious.
Currently the data is an a MySQL database, use by a PHP application; however, I'm open to other storage methods if they're a key part in accomplishing this. here.
There are a number of ways of determining the distance between two points, but for plotting points on a 2-D graph you probably want the Euclidean distance. If (x1, y1) represents your first point and (x2, y2) represents your second, the distance is
d = sqrt( (x2-x1)^2 + (y2-y1)^2 )
Regarding grouping, you may want to use some sort of 2-D mean to determine how "close" things are to each other. For example, if you have three points, (x1, y1), (x2, y2), (x3, y3), you can find the center of these three points by simple averaging:
x(mean) = (x1+x2+x3)/3
y(mean) = (y1+y2+y3)/3
You can then see how close each is to the center to determine whether it should be part of the "cluster".
There are a number of ways one can define clusters, all of which use some variant of a clustering algorithm. I'm in a rush now and don't have time to summarize, but check out the link and the algorithms, and hopefully other people will be able to provide more detail. Good luck!
Use something similar to the method you outlined in your question to get an approximate set of results, then whittle that approximate set down by doing proper calculations. If you pick your grid size (i.e. how much you round off your co-ordinates) correctly, you can at least hope to reduce the amount of work to be done to an acceptable level, although you have to manage what that grid size is.
For example, the earthdistance extension to PostgreSQL works by converting lat/long pairs to (x,y,z) cartesian co-ordinates, modelling the Earth as a uniform sphere. PostgreSQL has a sophisticated indexing system that allows these co-ordinates, or boxes around them, to be indexed into R-trees, but you can whack something together that is still useful without that.
If you take your (x,y,z) triple and round off- i.e. multiply by some factor and truncate to integer- you then have three integers that you can concatenate to produce a "box name", which identifies a box in your "grid" that the point is in.
If you want to search for all points within X km of some target point, you generate all the "box names" around that point (once you've converted your target point to an (x,y,z) triple as well, that's easy) and eliminate all the boxes that don't intersect the Earth's surface (tricker, but use of the x^2+y^2+z^2=R^2 formula at each corner will tell you) you end up with a list of boxes target points can be in- so just search for all points matching one of those boxes, which will also return you some extra points. So as a final stage you need to calculate the actual distance to your target point and eliminate some (again, this can be sped up by working in Cartesian co-ordinates and converting your target great-circle distance radius to secant distance).
The fiddling around comes down to making sure you don't have to search too many boxes, but at the same time don't bring in too many extra points. I've found it useful to index each point on several different grids (e.g. resolutions of 1Km, 5Km, 25Km, 125Km etc). Ideally you want to be searching just one box, remember it expands to at least 27 as soon as your target radius exceeds your grid size.
I've used this technique to construct a spatial index using Lucene rather than doing calculations in a SQL databases. It does work, although there is some fiddling to set it up, and the indices take a while to generate and are quite big. Using an R-tree to hold all the co-ordinates is a much nicer approach, but would take more custom coding- this technique basically just requires a fast hash-table lookup (so would probably work well with all the NoSQL databases that are the rage these days, and should be usable in a SQL database too).
Maybe overkill, but it seems to me a clustering problem: distance measure will determine how the similarity of two elements is calculated. If you need a less naive solution try Data Mining: Practical Machine Learning Tools and Techniques, and use Weka or Orange
If I were tackling it, I'd start with a grid. Put each point into a square on the grid. Look for grids that are densely populated. If the adjacent grids aren't populated, then you have a decent group.
If you have adjacent densely populated grids, you can always drop a circle at the center of each grid and optimize for circle area vs (number of points in the circle * some tunable weight). Not perfect, but easy. Better groupings are much more complicated optimization problems.
Facing a similar issue, I've just floor the Longitude and Latitude until I got the required 'closeness' in meters. In my case, floor to 4 digits got me locations grouped when they are approx. 13 meters apart.
If the Long or Lat are negatives - replace floor with ceil
First FLOOR (or CEIL) to required precision and then GROUP on the rounded long and lat.
The code to measure distance between two geo locations was borrowed from Getting distance between two points based on latitude/longitude
from math import sin, cos, sqrt, atan2, radians
R = 6373.0
lat1 = radians(48.71953)
lon1 = radians(-73.72882)
lat2 = radians(48.719)
lon2 = radians(-73.728)
dlon = lon2 - lon1
dlat = lat2 - lat1
a = sin(dlat / 2)**2 + cos(lat1) * cos(lat2) * sin(dlon / 2)**2
c = 2 * atan2(sqrt(a), sqrt(1 - a))
distance = (R * c)*1000
print("Distance in meters:", round(distance))
Distance in meters: 84
As expected, the distance is larger in the south, and smaller in the north - for the same angle.
For the same coordinates, but on the equator, the distance is 109 meters (modify the latitudes to 0.71953 and 0.719).
I modified the number of digits in the following and always kept one-click on Long and one on Lats, and measured the resulting distances:
lat1 = radians(48.71953)
lon1 = radians(-73.72882)
lat2 = radians(48.71954)
lon2 = radians(-73.72883)
Distance in meters 1
lat1 = radians(48.7195)
lon1 = radians(-73.7288)
lat2 = radians(48.7196)
lon2 = radians(-73.7289)
Distance in meters 13
lat1 = radians(48.719)
lon1 = radians(-73.728)
lat2 = radians(48.720)
lon2 = radians(-73.729)
Distance in meters 133
lat1 = radians(48.71)
lon1 = radians(-73.72)
lat2 = radians(48.72)
lon2 = radians(-73.73)
Distance in meters 1333
Summary: Floor / Ceil the longitude and latitude to 4 digits, will help you group on locations that are approximately 13 meters apart.
This number changes depending on the above equation: larger near the equator and smaller in the north.
If you are considering latitude and longitude there are several factors to be considered in real time data: obstructions, such as rivers and lakes, and facilities, such as bridges and tunnels. You cannot group them simply; if you use the simple algorithm as k means you will not be able to group them. I think you should go for the spatial clustering methods as partitioning CLARANS method.
I wanted to simulate the following through animation :
A ball starts with a certain velocity at the bottom most point of
a vertical circular loop and keeps rolling in it until its velocity permits.
For this, I wanted to find velocity/x/y vs. time equation.
For e.g. if the ball had mass : 5Kg, radius of the circular loop = 10m,
and initial velocity of the ball is 200 m/s, what will its velocity and (x,y) position
be after 5 seconds?
thanks.
Sliding, frictionless case with a point-particle ball
In this case we aren't worrying about rotational energy and are assuming that the ball is actually a point particle. Then, in order for the ball to stay on at the top, the centripetal force condition has to be satisfied:
m * v_top^2 / r = m * g
so
v_top = sqrt(r * g)
So the minimum initial velocity is determined by:
1 / 2 * m * v0^2 >= 1 / 2 * m * v_top^2 + m * g * 2 * r
v0 >= sqrt(5 * r * g)
This is similar to what Pete said, except that he forgot the centripetal force condition to stay on at the top.
Next, the acceleration tangential to the track is given by:
a = - g * sin(theta)
but a = r * alpha = r * d^2(theta)/dt^2 where alpha is the rotational acceleration. Thus, we get
r * d^2(theta)/dt^2 = g * sin(theta)
However, I don't know of an analytical solution to this differential equation and Mathematica was stumbling with finding one too. You can't just move the dts to the other side and integrate because theta is a function of t. I would recommend solving it by numerical means such as a Runga-Kutte or maybe the Verlet method. I solved it using Mathematica for the parameters you gave, but with the ball moving so quickly, it doesn't really slow down much in going around. When I lowered the initial velocity though, I was able to see the speeding up and slowing down by plotting theta as a function of time.
Adding in other things like a finite ball radius, rotational energy and friction are certainly doable, but I would worry about being able to solve this first case before moving on because it only gets more complicated from here. By the way, with the friction you will have to choose some kinetic coefficient of friction for your given materials which will of course be proportional to the normal force exerted on the ball by the track which can be solved for by summing the force components along the radius of the circle and don't forget to include the centripetal force condition.
If you haven't done this sort of physics before, I definitely recommend getting a introductory good book on physics (with calculus) and working through it. You only need to bother with the sections that apply to mechanics though that is a very large section of the book probably. There might be better routes to pursue though like some of the resources in this question.
If there are no acceleration (x,y) =(xstart+ vx*time ,ystart + vy*time) and speed remain the same, and it is not related to the radius
Since the velocity is constant you will have an angular velocity of omega = vel / radius. You will obtain how many radians you ball will move per second over its circular path.
To get the position at time t you just have to exploit polar coordinates:
x = x_center + sin( 3/2*PI + omega*t)*radius
y = y_center + cos( 3/2*PI + omega*t)*radius
This because you start from bottom point of the circle (so its 3/2*PI) plus how many radiants you move every second (we obtained it from tangential velocity). All multiplied for the radius, otherwise you will consider a unity circle.
EDIT: Since you wonder how to find a position of an object that is subject to many different forces I can tell you that usually a physical engine doesn't care about finding equations of moving objects. It just applies forces to objects considering their intended motions (like your circular one) or environmental factors (like gravity or friction) and calculates coordinates step by step by applying forces and using an integrator to see the results.
Ignoring friction, the forces on the ball are gravity and the track.
First, there are two main cases - is the velocity enough for the ball to loop-the-loop or not:
initial energy = 1/2 m v² = 0.5 * 5 * 200 * 200
potential energy = m g h = 5 * 9.8 * 20
so it will go round the whole loop.
Initially the ball is at the bottom of the loop, theta = 0
The acceleration on the ball is the component of g along the track
a = g⋅sin theta
The distance travelled is theta * radius. It is also the double integral of acceleration against time.
theta ⋅ radius = double integral of acceleration against time
Integrating acceleration once gives velocity, integrating velocity gives distance.
so solve this for t:
theta ⋅ r = ∫(∫ g⋅sin theta.dt).dt
then your x and y are trivial functions of theta.
Whether you solve it analytically or numerically is up to you.
With dynamic friction, friction is usually proportional to the normal force on the bodies. So this will equal the centripetal force - proportional to the square of the angular velocity, and the component of gravity normal to the track (g sin theta)
You didn't tell anything about how you want your velocity to change. Do you have any friction model? If there is no friction, then the formulas are simple:
length = velocity*t
x = sin(length)*radius
y = -cos(length)*radius
If the velocity is changing, then you have to change length to something like
length = integral over dt[0..t] (velocity dt)
The only thing I wanted to add is the if this is real ball (sphere) with mass 5kg then it must have a diameter dia=(6*m/(PI*rho))^(1/3) where rho is the density of the material. For steel (rho=7680) the diameter is dia=0.1075 meters. Therefore the pitch radius (radius at which the center of gravity of the ball rides on) is equal to R=10-(dia/2) or R=9.9466 meters.
The problem gets a little more complex when friction is included. For one you have to consider the direction of friction (assuming dry friction theory). That depends on the amount the ball rotates in its axis and that depends on moment of inertia of the ball.
When you do the simulation you might want to monitor the total kinetic energy + the total potential energy and make sure your are not adding energy to the system (or taking away). [Don't forget to include the rotational component for the kinetic energy]
Get a standard book on dynamics, and I am sure a similar problem is already described in the book.I would recommend "Vector Mechanic for Engineers - Dynamics".