Generating a 2D point cloud map

Hello there

I have to generate point clouds which represents a room or a floor, as example. Also it should be able to add an error to the generator of the point cloud to make it more real (like a noise).

With a small section of the map it should than be possible to find out the location of this small section. Like SLAM without the mapping. Just localisation.

Maybe someone has a solution or some ideas to solve this.

Kind regard

Ramon

[darova]

How does the cloud look like?

[Ramon]

something like that....in the first picture there are multiple measurments overlapped and in the second picture is one measurment.

I think it's possible to make it somehow with linspace and additional errors .

the restriction later is, that the maze just have angles of 90°... maybe this makes it easier.

[darova]

You have black line and want red points? Is it correct?

[Ramon]

yes, but it has to be something like a normal distribution...like 68% are on the line and the others are offset to the black line....something like that

[darova]

Any ideas how to implement it?

[Ramon]

Yes, i made some functions to make lines with errors.

%% Functions
% Function for generating a line out of points with an additional random
% error for one axis 
function[v_ret] = calc_v(start,fin,N,error)
    dv = (fin-start)/N;
    v = linspace(start,fin,N);
        if dv == 0
            dv = 1;
        end
    v_ret = v-(-error + (error+error)*rand(1,N))*dv;
end

% Function to concatenate two arrays. Returns concatenated array.
function[arr_x,arr_y] = concator(array_x,extension_x,array_y,extension_y)
    arr_x = [array_x extension_x];
    arr_y = [array_y extension_y];
end

% Function to add a cos-error 
function[arrayWithError] = addCosError(arr,step,err_amp,err_step)
    format = size(arr(1:step:end));
    len = format(2);
    i = linspace(1,err_step,len);
    arr(1:step:end) = arr(1:step:end) + err_amp*cos(pi.*i);
    arrayWithError = arr;

end

% Function for generate a point line for mapping
function [arr_x,arr_y] = setLine(x_start,x_fin,y_start,y_fin,rand_error,step_to_add_error,err_amp,err_period)
if (x_fin - x_start) > (y_fin - y_start)
    N = x_fin - x_start;
else 
    N = y_fin - y_start;
end
arr_x = addCosError(calc_v(x_start,x_fin,N,rand_error),step_to_add_error,err_amp,err_period);
arr_y = addCosError(calc_v(y_start,y_fin,N,rand_error),step_to_add_error,err_amp,err_period);
end

The result is like this:

An example map without errors.

And the example map with errors.

Answer 1

[Ramon]

The Answer is here: https://ch.mathworks.com/matlabcentral/answers/482095-generating-a-2d-point-cloud-map#comment_750814

[darova]

The best solution is the solution made by yourself :)