Blog posts

Under construction....


Why is RRT efficient?

less than 1 minute read


Trajectory optimization is a local method, it only explores the neighborhood of an initial seed, besides, in the direct transcription, we discretize along the trajectory instead of every dimension of the C-space, thus it scales well in high dimensional space. But RRT is a global method, what makes it efficient in exploring a collision-free path in high dimensional space?

Convex optimization, QP, SQP

3 minute read


Definition of convex function

For all ${ 0\leq t \leq 1}$ and all ${ x_{1},x_{2}\in X,}$ ${ f\left(tx_{1}+(1-t)x_{2}\right)~\leq ~tf\left(x_{1}\right)+(1-t)f\left(x_{2}\right).}$

Quadratic forms and Convexity

Quadratic form in variables ${ x_1,x_2…, x_n}$ is a polynomial function $Q$, where all the terms in $Q(x_1, x_2,…, x_n)$ have order two. Quadratic functions $\neq$ convex functions.


Simulated Annealing

less than 1 minute read


Simulated Annealing is a probabilistic method for approximating the global optimum of a given function. It is helpful especially in the case of large search space.

  • Make use of randomness, random walk on a search graph.
  • Transition probabilities..
  • Higher probability of accepting worse solutions in the begining (high temperature).