question about ScanMatcher::score

[asmashine]

in this function ,to caculate pfree,there are
double freeDelta=map.getDelta()*m_freeCellRatio;and
pfree.x+=(*r-map.getDelta()*freeDelta)*cos(lp.theta+*angle); pfree.y+=(*r-map.getDelta()*freeDelta)*sin(lp.theta+*angle);;
obviously r - delta^2sqrt(2) (freeDelta = sqrt(2))
i think r - deltasqart(2) can show the range of pfree ,but the equation above i don't know its mean.

[lmy19880626]

me too! I need the explanation.

[KOTOKORURU]

I have the same question ,have you sloved it？

[1344618323]

me too! I think there's a bug!

[topin89]

I think this is a bug that was shadowed by another bug, #25
making everything about pfree, ipfree, pf and fcell irrelevant.

And yes,

pfree.x+=(*r-map.getDelta()*freeDelta)*cos(lp.theta+*angle);
pfree.y+=(*r-map.getDelta()*freeDelta)*sin(lp.theta+*angle);

should be replaced with

pfree.x+=(*r-*freeDelta)*cos(lp.theta+*angle);
pfree.y+=(*r-*freeDelta)*sin(lp.theta+*angle);

With these two bugs solved, gmapping stops looking for best score inside a wall.

[topin89]

Oh, there is a PR to fix that: #10