--- title: 'Trajectory Trees: Hierarchical Motion & Data Structures' url: https://www.emergentmind.com/topics/trajectory-trees type: topic --- # Trajectory Trees: Hierarchical Motion & Data Structures Trajectory trees are tree-structured representations in which trajectories, trajectory segments, or trajectory-derived keys are organized hierarchically. In the literature, the term does not denote a single canonical object. In kinodynamic planning, it refers to search trees whose nodes are system states and whose edges are dynamically feasible trajectory segments [2109.07741]. In uncertainty-aware planning, it may denote one Monte Carlo Tree Search tree per sampled start state, or a pair of aligned trees for ego actions and environment responses [2203.04452] [2301.11902]. In prediction, a root-to-leaf path can encode a multimodal future, a recursive sequence of sub-goals, or an interpretable intention hypothesis [2205.13296] [1906.05329] [2511.14265]. In data management, trajectory trees appear as hierarchical indexes, compressed snapshot structures, or tries over trajectory sketches [2406.05327] [1611.05247] [2005.10917]. The common feature is hierarchical branching over motion, but the semantics of nodes, edges, and branching vary substantially by domain. ## 1. Structural interpretations across domains Across the surveyed work, trajectory trees serve at least four distinct roles: executable motion search, contingent decision structure, multimodal prediction scaffold, and data-management index. This plurality is explicit in the formal definitions used by the papers, rather than being a matter of informal terminology [2109.07741] [2203.04452] [2301.11902] [2205.13296] [2406.05327]. | Setting | What the tree stores | Meaning of branching | |---|---|---| | Kinodynamic planning | States and dynamically feasible trajectory segments | Alternative expansions or local deformations | | Uncertainty-aware planning | Belief-sample trees, scenario trees, or observation-contingent branches | Distinct hypotheses, sampled start states, or future responses | | Prediction | Motion primitives, sub-goals, or intention prototypes | Multimodal future behaviors | | Indexing and retrieval | Bounding predicates, snapshots, compressed logs, or sketch prefixes | Hierarchical partitioning for pruning and lookup | In kinodynamic motion planning, the tree is a state-space search structure. In "STD-Trees: Spatio-temporal Deformable Trees for Multirotors Kinodynamic Planning" [2109.07741], nodes are multirotor states and edges are LQMT trajectory segments with associated control and duration. In "Train-Once Plan-Anywhere Kinodynamic Motion Planning via Diffusion Trees" [2508.21001], edges are short propagated subtrajectories produced by applying a learned action sequence to the true dynamics. In "Model Predictive Trees: Sample-Efficient Receding Horizon Planning with Reusable Tree Search" [2411.15651], a finite-horizon UCT tree is rebuilt and re-rooted at each MPC iteration. In uncertainty-aware planning, the tree may no longer be a single forward rollout of physical state. In "Cooperative Trajectory Planning in Uncertain Environments with Monte Carlo Tree Search and Risk Metrics" [2203.04452], a trajectory tree is the MCTS tree rooted at one sampled start state from a root belief, and the planner builds multiple such trees. In "Tree-structured Policy Planning with Learned Behavior Models" [2301.11902], an ego trajectory tree and a scenario tree are coupled stage by stage. In "Optimizing Trajectory-Trees in Belief Space" [2605.01860], a trajectory-tree alternates deterministic trajectory stages and probabilistic belief-update stages. Prediction papers use trees differently. In "Social Interpretable Tree for Pedestrian Trajectory Prediction" [2205.13296], a root-to-leaf path corresponds to a coarse candidate future defined by motion primitives such as forward, left-turn, and right-turn. In the sub-goal-tree framework [1906.05329] [2002.12361], internal nodes recursively split a start-goal segment by predicting midpoints. In "Unified Multimodal Vessel Trajectory Prediction with Explainable Navigation Intention" [2511.14265], branches correspond to sustained intention prototypes and leaves arise from transient latent sampling. Database papers treat trajectory trees as indexes rather than plans. "Multi-Entry Generalized Search Trees for Indexing Trajectories" [2406.05327] uses multi-entry GiST and SP-GiST structures, "A new method to index and store spatio-temporal data" [1611.05247] uses a $k^2$-tree over spatial snapshots plus compressed movement logs, "GraCT" [1612.03308] combines $k^2$-tree snapshots with grammar-compressed logs, and "Succinct Trit-array Trie for Scalable Trajectory Similarity Search" [2005.10917] organizes LSH sketches of trajectories in succinct tries. ## 2. Kinodynamic planning and control In kinodynamic planning, trajectory trees are central because they simultaneously encode reachability, dynamics, and cost. STD-Trees formalize this explicitly through deformation units consisting of one node and all incident edges, with decision variables given by the node state $\mathbf{x}_n$ and the durations of the incoming and outgoing edges $\mathbf{T}_n$ [2109.07741]. The local objective is a weighted sum of edge costs over the corresponding sub-tree, using path-count weights $d_i$, plus soft penalties for obstacle avoidance and dynamical limits. For multirotors, the paper uses differential flatness, a third-order integrator model with $s=3$, jerk control, and polynomial edge parameterization. A notable feature is that optimization is both spatial and temporal: the node location and the edge durations are adjusted together, and the resulting non-smooth problem is solved with LMBM because distance-field interpolation is non-smooth [2109.07741]. Algorithmically, STD-Trees insert deformation into an RRT-style loop after the first feasible solution is found. The paper defines NODE, TRUNK, BRANCH, and TREE variants, followed by cascade rewiring in the style of RRT$^\#$ [2109.07741]. The reported results show that all variants converge much faster than the no-deformation baseline, that BRANCH slightly outperforms TREE, and that spatio-temporal deformation outperforms spatial-only deformation across kRRT, kRRT*, and kRRT$^\#$ [2109.07741]. The paper further argues that local deformation preserves the current homotopy class of a root-to-leaf path unless the optimized state opens a better local corridor within that class. Diffusion-guided tree growth replaces local optimization with learned action sampling. DiTree plugs a conditional diffusion or flow-matching policy into the action-selection step of an RRT-like planner and samples short action sequences conditioned on the expansion node, a relative target, and a local occupancy grid [2508.21001]. The tree remains a standard kinodynamic search tree: forward propagation uses the given dynamics, and every edge is accepted only after collision checking. The paper proves probabilistic completeness under Lipschitz dynamics and a full-support action sampler, with failure probability bounded by $a e^{-bk}$, and safety follows from collision checking of every propagated edge [2508.21001]. Empirically, DiTree is reported as on average $3\times$ faster than classical SBPs and as achieving roughly $30\%$ higher success rate across out-of-distribution scenarios [2508.21001]. Model Predictive Trees shift emphasis from expansion quality to search reuse. MPT grows a depth-$K$ UCT tree, selects the best child of the root, executes one step, and then reuses the entire selected subtree as the next MPC warm start [2411.15651]. The node state includes cumulative value and visit count, and the UCT score is $$ \textup{UCT(c)}=\frac{\text{c}.V}{\text{c}.N}+\epsilon\sqrt{\frac{\log(\text{p}.N)}{\text{c}.N}}. $$ The paper derives a steady-state tracking-error bound under time-varying dynamics, $$ \|x_\infty-x^d_\infty\|\le \sqrt{\frac{\overline{m}}{\underline{m}}}\frac{(K+1)\eta+\varepsilon}{1-\alpha}, $$ which makes the reuse trade-off explicit: larger $K$ improves foresight but worsens the bound when the disturbance changes quickly [2411.15651]. Numerical studies report that MPT reaches average cumulative reward $V=80$ with $L=180$ simulated trajectories, compared with $5200$ for CEM, $3000$ for CEM-Reuse, and more than $30000$ for UCT; onboard hardware experiments run at $5$ Hz with about $2100$ simulated trajectories every $0.2$ s [2411.15651]. ## 3. Uncertainty, interaction, and risk-sensitive branching When uncertainty is central, trajectory trees often branch over hypotheses rather than controls alone. In the cooperative MCTS planner of [2203.04452], the world state is only known through a root belief $b_0$ built from Gaussian, unbiased, independent sensor models with diagonal covariance. Multiple start states are sampled from that belief, progressive widening decides when to add new start states, and one continuous-action MCTS tree is built for each sampled start state. The resulting action values are then aggregated across trees by kernel regression using a Gaussian radial basis kernel, and final selection is performed by KRLCB or CCVaR [2203.04452]. This produces a two-level interpretation of a trajectory tree: locally, each tree is a determinized MCTS search tree; globally, the collection of trees approximates a return distribution over uncertain initial conditions. On 15 scenarios with up to 8 vehicles and 300 random seeds, the uncertain baseline degrades to mean success rates of roughly $0.2$–$0.3$, whereas KRLCB and CVaR rise to around $0.5$–$0.55$, with modest runtime overheads [2203.04452]. Tree Policy Planning makes the contingency structure explicit by separating ego choices from environment responses [2301.11902]. The ego trajectory tree contains dynamically feasible stagewise motion segments generated from sampled terminal states and cubic polynomials, while the scenario tree contains stagewise multimodal, ego-conditioned environment futures with branch probabilities from a learned predictor. Because both trees share the same stage structure, the planner reduces to a finite-horizon discrete MDP with Bellman backup $$ V(r^i,e^i)=L_i(r^i,e^i)+\min_{r^{i+1}\in \text{Ch}(r^i)}\mathbb{E}_{e^{i+1}}\big[V(r^{i+1},e^{i+1})\mid e^i\big]. $$ The ego-conditioned version requires a causally consistent ensemble of scenario trees [2301.11902]. In closed-loop nuScenes simulations, TPP reduces crash and offroad rates relative to non-contingency baselines across rasterized, PredictionNet, and Agentformer predictors, although temporally inconsistent Gaussian latents degrade performance [2301.11902]. Belief-space trajectory trees push this logic further by optimizing branches directly in belief space [2605.01860]. In PO-MPC, there is a single branching horizon: controls before branching are constrained by non-anticipativity, and each branch is weighted by its hypothesis probability $p(m)$. In PO-LGP, trajectory-trees at the motion level are paired with decision trees at the task level, with deterministic observations pruning hypotheses [2605.01860]. The generic PO-MPC problem enforces shared pre-branch controls, $$ z_m(t)=\tilde z(t),\quad \forall m,\ \forall t