This GIT respository contains all files needed for an adequate analysis of the gait (6MWT) accelerometer data.
You can not select more than 25 topics Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.
 
 
 
 

1 lines
2.7 KiB

function Divergence=div_calc(state,ws_sec,fs,period_sec,progress) % calculate divergence curve (needed for max lyapunov exponent). % Input: % state: state space % ws_sec: window size over which divergence will be calculated(in seconds) % fs: sample frequency % period_sec: indicates period of time-wise near samples to exclude in % nearest neighbour search % progress: show progress or not % % Output: % Divergence: the divergence curve % % History: % August 2011: v1, Sietse Rispens based on version KvS en SMB % 1 November 2012, Sietse Rispens: Use div_calc_shorttimeseries for short % time series if size(state,1) <= 10000 Divergence=div_calc_shorttimeseries(state,ws_sec,fs,period_sec,progress); return; end [N,D]=size(state); ws=round(ws_sec*fs); Period=round(period_sec*fs); if N<ws, error('ws shall not be larger than N'); end NCompleteWindow = N-ws+1; Divergence_sum=zeros(1,ws); Divergence_count=zeros(1,ws); SumStd = sqrt(sum(std(state,1).^2)); DistLimStart = SumStd*nthroot(1/NCompleteWindow,D); DistLim = DistLimStart; TreeRoot=kdtree_build(state(1:NCompleteWindow,:)); k0 = 2; % The initial number of nearest neighbors to look for kmax = min(2*Period + 2, NCompleteWindow); % The maximum number of nearest neighbors to look for for i = 1:NCompleteWindow StateI = state(i,:); if ~isnan(StateI) DistLim = DistLim*nthroot(1/5,D); DistLim = min(DistLim,DistLimStart); Index = []; k=k0; kmaxreached = 0; while isempty(Index) && ~kmaxreached [Idx, Dist] = kdtree_k_nearest_neighbors(TreeRoot,StateI,k); Dist(abs(i-Idx) < Period)=nan; Index = Idx(find(Dist==min(Dist),1,'first')); if k >= kmax kmaxreached = 1; elseif isempty(Index) k = min(kmax,k*2); end end if ~isempty(Index) if i+ws>N || Index+ws>N % do not use these data else if ~isempty(Index) DistCurve = log(sqrt(sum((state(i:i+ws-1,:)-state(Index:Index+ws-1,:)).^2,2))); NotNan = ~isnan(DistCurve); Divergence_sum(NotNan) = Divergence_sum(NotNan) + DistCurve(NotNan)'; Divergence_count(NotNan) = Divergence_count(NotNan) + 1; end end end end if progress > 0 if mod(i,round(progress))==0 if i>round(progress) fprintf('\b\b\b\b\b\b\b\b\b\b'); end fprintf('i=%8d', i); end end end kdtree_delete(TreeRoot); % Free the pointer to k-d-tree if progress > 0 fprintf('\n'); end Divergence= (Divergence_sum./Divergence_count);