Smooth Spatiotemporal Tube Synthesis for Prescribed-Time Reach-Avoid-Stay Control

Basic Information

• Paper ID: 2510.11583
• Title: Smooth Spatiotemporal Tube Synthesis for Prescribed-Time Reach-Avoid-Stay Control
• Authors: Siddhartha Upadhyay, Ratnangshu Das, Pushpak Jagtap (Indian Institute of Science)
• Classification: eess.SY (Systems and Control), cs.RO (Robotics), cs.SY (Systems and Control)
• Publication Date: October 14, 2025
• Paper Link: https://arxiv.org/abs/2510.11583

Abstract

This paper addresses the controller synthesis problem for control-affine nonlinear systems to satisfy prescribed-time Reach-Avoid-Stay (RAS) specifications. The research aims to improve existing spatiotemporal tube (STT) based methods by eliminating detour functions that cause abrupt tube variations and high control costs. An adaptive framework is proposed to construct smooth STTs around static unsafe sets, enabling continuous obstacle avoidance while guiding the system to reach the target within prescribed time. Closed-form, approximation-free control laws are derived to ensure system trajectories remain within the tube and satisfy RAS tasks. Case studies validate the method's effectiveness, demonstrating significantly reduced control costs compared to previous approaches.

Research Background and Motivation

1. Problem Definition

The core problem addressed in this paper is designing controllers for control-affine nonlinear systems to satisfy prescribed-time Reach-Avoid-Stay (RAS) specifications. RAS tasks require the system to depart from an initial set, reach a target set within prescribed time, continuously avoid unsafe regions, and remain in the target region after arrival.

2. Problem Significance

RAS properties are fundamental components for defining more complex temporal logic specifications and are crucial for safety-critical systems. They support the development of robust control strategies ensuring system states remain in safe regions, with broad applications in robot navigation, autonomous driving, aerospace, and related fields.

3. Limitations of Existing Methods

Current approaches suffer from the following limitations:

• Abstraction-based methods: Symbolic control methods based on state space discretization face the curse of dimensionality
• Optimization-based methods: Control barrier function approaches are computationally expensive and difficult to scale to high-dimensional systems
• Funnel control: While computationally efficient, handling convex specifications (such as collision avoidance constraints) remains challenging
• Existing STT methods: Use of detour functions causes abrupt tube variations, resulting in high control costs

4. Research Motivation

Based on these limitations, this paper proposes a smooth STT synthesis method to:

• Eliminate the use of detour functions and avoid abrupt tube variations
• Reduce control costs and improve system performance
• Maintain satisfaction guarantees for prescribed-time RAS specifications
• Provide closed-form, approximation-free control solutions

Core Contributions

1. Proposed smooth STT adaptive framework: Designed a novel adaptive framework capable of constructing smooth spatiotemporal tubes in real-time, avoiding the use of detour functions
2. Eliminated abrupt tube variations: Through smooth continuous deformation for obstacle avoidance constraints, significantly reducing control costs
3. Maintained theoretical guarantees: While reducing control costs, preserved theoretical guarantees for prescribed-time RAS specification satisfaction
4. Provided closed-form control laws: Derived approximation-free closed-form control rules applicable to control-affine systems with unknown dynamics
5. Validated method effectiveness: Verified practical utility and superiority through simulation case studies of omnidirectional mobile robots

Methodology Details

Task Definition

Prescribed-time RAS task: Given unsafe set U, initial set S⊂ℝⁿ\U, and target set T⊂ℝⁿ\U, for any initial state x(0)∈S, there exists time t∈0,tc such that x(t)∈T, with trajectories avoiding the unsafe set at all times, i.e., x(s)∈ℝⁿ\U for all s∈0,tc, where tc∈ℝ⁺ is the prescribed completion time.

System Model

Consider the control-affine nonlinear system:

S: ẋ = f(x) + g(x)u + w

where:

• x(t)∈ℝⁿ: state vector
• u(t)∈ℝⁿ: control input vector
• w(t)∈W⊂ℝⁿ: unknown bounded disturbance
• f(x) and g(x): unknown locally Lipschitz functions

Model Architecture

1. Reachability Tube Design

Define a hyperrectangle centered at initial state x(0):

Ŝ := ∏[i=1,n] [xi(0) - di,S, xi(0) + di,S] ⊂ S

Reachability tube boundary ρ(t) evolves according to the differential equation:

ρ̇i(t) = {
  tc(T̂i - Ŝi)/((tc-t)²) sech²(t/(tc-t)), if t < tc
  0, if t ≥ tc
}

2. Obstacle Avoidance Specification Handling

Unsafe set modeled as: U = ⋃j∈1,nu U⁽ʲ⁾ ⊂ ℝⁿ

By computing the intersection time interval t⁽ʲ⁾, t̄⁽ʲ⁾ of the tube with the j-th unsafe set:

tj = (max[i∈[1,n]] min āi) tc, t̄j = (min[i∈[1,n]] max āi) tc

3. Adaptive Tube Correction

Propose an adaptive framework to correct STT:

γ̇i,L(t) = {
  ρ̇i(t), if i ≠ k
  α₁⁽ʲ⁾ρ̇i(t) + α₂⁽ʲ⁾ϕ₁⁽ʲ⁾(t) + α₃⁽ʲ⁾ϕ₂⁽ʲ⁾(t), if i = k
}

where weight functions α₁⁽ʲ⁾, α₂⁽ʲ⁾, α₃⁽ʲ⁾ are defined based on s(t) = 0.5tanh(t/v), implementing three-stage smooth switching.

Technical Innovations

1. Elimination of Detour Functions

Unlike existing methods, this paper avoids using detour functions that cause abrupt variations, instead implementing smooth tube deformation through adaptive weight functions.

2. Three-Stage Adaptive Mechanism

• Stage 1: α₁⁽ʲ⁾=1, tube directed toward target set
• Stage 2: α₂⁽ʲ⁾=1, tube circumvents unsafe set projection
• Stage 3: α₃⁽ʲ⁾=1, tube smoothly returns to original reachability boundary

3. Closed-Form Control Law

Derive approximation-free control law:

u(x,t) = -κξ(x,t)ε(x,t)

where ξ(x,t) and ε(x,t) are defined based on normalized error and transformation error.

Experimental Setup

System Model

Employs omnidirectional mobile robot dynamics:

[ẋ₁; ẋ₂; ẋ₃] = [cos x₃ -sin x₃ 0; sin x₃ cos x₃ 0; 0 0 1][v₁; v₂; ω] + w(t)

Scenario Configuration

• Initial region: S = 0, 0.5 × 0, 0.5
• Target region: T = 11, 11.5 × 7, 7.5
• Obstacles:
  • U₁ = 1.5, 2 × 0.5, 3
  • U₂ = 5.2, 6.8 × 3.2, 4
  • U₃ = 7, 8 × 0, 8
• Prescribed time: tc = 80s

Evaluation Metrics

Primary evaluation focuses on the degree of control cost reduction through comparison with detour function-based methods.

Experimental Results

Main Results

1. Trajectory performance: System successfully reaches target region from initial region within prescribed time while avoiding all obstacles
2. Control cost: Significantly reduced control costs compared to detour function-based methods
3. Smoothness: STT design is smoother, avoiding abrupt variations

Case Analysis

Figure 2 displays robot navigation trajectories in 2D environment, with black solid lines representing system trajectories successfully avoiding all red obstacles and reaching the green target region. Figure 3 comparison results clearly demonstrate the significant advantages of the proposed method in control cost reduction.

Theoretical Verification

Theorem 3.3 proves that the constructed STT satisfies the following conditions:

• Tube originates from initial set: Γ(0) ⊂ S
• Reaches target within prescribed time: Γ(tc) ⊂ T
• Avoids unsafe set throughout: Γ(t) ∩ U = ∅
• Tube boundaries satisfy constraints

Related Work

Main Research Directions

1. Abstraction-based methods: Symbolic control synthesis through state space discretization
2. Barrier function methods: Optimization-based approaches using control Lyapunov-barrier functions
3. Funnel control: Efficient tracking control methods
4. STT methods: Development of spatiotemporal tube frameworks

Advantages of This Work

Compared to existing work, the proposed method:

• Avoids the curse of dimensionality
• Reduces computational complexity
• Eliminates abrupt tube variations
• Provides closed-form solutions

Conclusions and Discussion

Main Conclusions

1. The proposed smooth STT synthesis method effectively solves prescribed-time RAS control problems
2. Eliminating detour functions significantly reduces control costs
3. The adaptive framework achieves smooth obstacle avoidance handling
4. Theoretical analysis guarantees method correctness and completeness

Limitations

1. Static obstacle assumption: Current method assumes unsafe sets are static and known
2. System class restrictions: Primarily targets control-affine systems
3. Temporal separation assumption: Requires unsafe sets to be sufficiently separated in time
4. High-dimensional scalability: Computational complexity for high-dimensional systems requires further verification

Future Directions

1. Extension to dynamic environments with time-varying obstacles
2. Generalization to broader nonlinear system classes
3. Coordinated control for multi-agent systems
4. Robustness analysis under stochastic disturbances

In-Depth Evaluation

Strengths

1. Strong theoretical innovation: The proposed smooth STT method demonstrates clear theoretical innovation by eliminating detour functions
2. High practical value: Significant control cost reduction has important implications for practical applications
3. Mathematical rigor: Provides comprehensive theoretical analysis and proofs
4. Method generality: Applicable to control-affine systems with unknown dynamics

Weaknesses

1. Limited experimental validation: Only provides one 2D robot navigation case study, lacking diverse verification
2. Strong assumptions: Static obstacles and temporal separation assumptions may be restrictive in practical applications
3. Insufficient computational complexity analysis: Lacks detailed analysis of method computational complexity
4. Insufficient robustness verification: Disturbance and uncertainty handling requires more validation

Impact

1. Academic contribution: Provides new ideas and technical approaches for STT method development
2. Application prospects: Shows good application potential in robot navigation, autonomous driving, and related fields
3. Method extensibility: Provides solid theoretical foundation for subsequent research

Applicable Scenarios

1. Robot path planning: Particularly suitable for robot navigation in known static environments
2. Unmanned aerial vehicle control: Applicable to UAV collision avoidance flight control
3. Industrial automation: Suitable for industrial robot systems requiring precise time control
4. Autonomous driving: Can be applied to autonomous vehicle control in structured environments

References

The paper cites 18 relevant references covering symbolic control, barrier function methods, funnel control, STT frameworks, and other related domains, providing solid theoretical foundation and comparison benchmarks for this research.